feat: character system v2 — schema upgrade, memory system, per-character TTS routing

Character schema v2: background, dialogue_style, appearance, skills, gaze_presets
with automatic v1→v2 migration. LLM-assisted character creation via Character MCP
server. Two-tier memory system (personal per-character + general shared) with
budget-based injection into LLM system prompt. Per-character TTS voice routing via
state file — Wyoming TTS server reads active config to route between Kokoro (local)
and ElevenLabs (cloud PCM 24kHz). Dashboard: memories page, conversation history,
character profile on cards, auto-TTS engine selection from character config.
Also includes VTube Studio expression bridge and ComfyUI API guide.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.env.example

@@ -9,6 +9,7 @@ OPENAI_API_KEY=
 DEEPSEEK_API_KEY=
 GEMINI_API_KEY=
 ELEVENLABS_API_KEY=
+GAZE_API_KEY=
 
 # ─── Data & Paths ──────────────────────────────────────────────────────────────
 DATA_DIR=${HOME}/homeai-data
@@ -40,10 +41,14 @@ OPEN_WEBUI_URL=http://localhost:3030
 OLLAMA_PRIMARY_MODEL=llama3.3:70b
 OLLAMA_FAST_MODEL=qwen2.5:7b
 
+# Medium model kept warm for voice pipeline (override per persona)
+# Used by preload-models.sh keep-warm daemon
+HOMEAI_MEDIUM_MODEL=qwen3.5:35b-a3b
+
 # ─── P3: Voice ─────────────────────────────────────────────────────────────────
 WYOMING_STT_URL=tcp://localhost:10300
 WYOMING_TTS_URL=tcp://localhost:10301
-ELEVENLABS_API_KEY=              # Create at elevenlabs.io if using elevenlabs TTS engine
+# ELEVENLABS_API_KEY is set above in API Keys section
 
 # ─── P4: Agent ─────────────────────────────────────────────────────────────────
 OPENCLAW_URL=http://localhost:8080

CLAUDE.md

@@ -26,6 +26,7 @@ All AI inference runs locally on this machine. No cloud dependency required (clo
 
 ### AI & LLM
 - **Ollama** — local LLM runtime (target models: Llama 3.3 70B, Qwen 2.5 72B)
+- **Model keep-warm daemon** — `preload-models.sh` runs as a loop, checks every 5 min, re-pins evicted models with `keep_alive=-1`. Keeps `qwen2.5:7b` (small/fast) and `$HOMEAI_MEDIUM_MODEL` (default: `qwen3.5:35b-a3b`) always loaded in VRAM. Medium model is configurable via env var for per-persona model assignment.
 - **Open WebUI** — browser-based chat interface, runs as Docker container
 
 ### Image Generation
@@ -35,7 +36,8 @@ All AI inference runs locally on this machine. No cloud dependency required (clo
 
 ### Speech
 - **Whisper.cpp** — speech-to-text, optimised for Apple Silicon/Neural Engine
-- **Kokoro TTS** — fast, lightweight text-to-speech (primary, low-latency)
+- **Kokoro TTS** — fast, lightweight text-to-speech (primary, low-latency, local)
+- **ElevenLabs TTS** — cloud voice cloning/synthesis (per-character voice ID, routed via state file)
 - **Chatterbox TTS** — voice cloning engine (Apple Silicon MPS optimised)
 - **Qwen3-TTS** — alternative voice cloning via MLX
 - **openWakeWord** — always-on wake word detection
@@ -49,11 +51,13 @@ All AI inference runs locally on this machine. No cloud dependency required (clo
 ### AI Agent / Orchestration
 - **OpenClaw** — primary AI agent layer; receives voice commands, calls tools, manages personality
 - **n8n** — visual workflow automation (Docker), chains AI actions
-- **mem0** — long-term memory layer for the AI character
+- **Character Memory System** — two-tier JSON-based memories (personal per-character + general shared), injected into LLM system prompt with budget truncation
 
 ### Character & Personality
-- **Character Manager** (built — see `character-manager.jsx`) — single config UI for personality, prompts, models, Live2D mappings, and notes
+- **Character Schema v2** — JSON spec with background, dialogue_style, appearance, skills, gaze_presets (v1 auto-migrated)
-- Character config exports to JSON, consumed by OpenClaw system prompt and pipeline
+- **HomeAI Dashboard** — unified web app: character editor, chat, memory manager, service dashboard
+- **Character MCP Server** — LLM-assisted character creation via Fandom wiki/Wikipedia lookup (Docker)
+- Character config stored as JSON files in `~/homeai-data/characters/`, consumed by bridge for system prompt construction
 
 ### Visual Representation
 - **VTube Studio** — Live2D model display on desktop (macOS) and mobile (iOS/Android)
@@ -85,47 +89,79 @@ All AI inference runs locally on this machine. No cloud dependency required (clo
 ESP32-S3-BOX-3 (room)
   → Wake word detected (openWakeWord, runs locally on device or Mac Mini)
   → Audio streamed to Mac Mini via Wyoming Satellite
-  → Whisper.cpp transcribes speech to text
+  → Whisper MLX transcribes speech to text
-  → OpenClaw receives text + context
+  → HA conversation agent → OpenClaw HTTP Bridge
-  → Ollama LLM generates response (with character persona from system prompt)
+  → Bridge resolves character (satellite_id → character mapping)
-  → mem0 updates long-term memory
+  → Bridge builds system prompt (profile + memories) and writes TTS config to state file
+  → OpenClaw CLI → Ollama LLM generates response
   → Response dispatched:
-      → Kokoro/Chatterbox renders TTS audio
+      → Wyoming TTS reads state file → routes to Kokoro (local) or ElevenLabs (cloud)
       → Audio sent back to ESP32-S3-BOX-3 (spoken response)
       → VTube Studio API triggered (expression + lip sync on desktop/mobile)
       → Home Assistant action called if applicable (lights, music, etc.)
 ```
 
+### Timeout Strategy
+
+The HTTP bridge checks Ollama `/api/ps` before each request to determine if the LLM is already loaded:
+
+| Layer | Warm (model loaded) | Cold (model loading) |
+|---|---|---|
+| HA conversation component | 200s | 200s |
+| OpenClaw HTTP bridge | 60s | 180s |
+| OpenClaw agent | 60s | 60s |
+
+The keep-warm daemon ensures models stay loaded, so cold starts should be rare (only after Ollama restarts or VRAM pressure).
+
 ---
 
 ## Character System
 
-The AI assistant has a defined personality managed via the Character Manager tool.
+The AI assistant has a defined personality managed via the HomeAI Dashboard (character editor + memory manager).
 
-Key config surfaces:
+### Character Schema v2
-- **System prompt** — injected into every Ollama request
+
-- **Voice clone reference** — `.wav` file path for Chatterbox/Qwen3-TTS
+Each character is a JSON file in `~/homeai-data/characters/` with:
-- **Live2D expression mappings** — idle, speaking, thinking, happy, error states
+- **System prompt** — core personality, injected into every LLM request
-- **VTube Studio WebSocket triggers** — JSON map of events to expressions
+- **Profile fields** — background, appearance, dialogue_style, skills array
+- **TTS config** — engine (kokoro/elevenlabs), kokoro_voice, elevenlabs_voice_id, elevenlabs_model, speed
+- **GAZE presets** — array of `{preset, trigger}` for image generation styles
 - **Custom prompt rules** — trigger/response overrides for specific contexts
-- **mem0** — persistent memory that evolves over time
 
-Character config JSON (exported from Character Manager) is the single source of truth consumed by all pipeline components.
+### Memory System
+
+Two-tier memory stored as JSON in `~/homeai-data/memories/`:
+- **Personal memories** (`personal/{character_id}.json`) — per-character, about user interactions
+- **General memories** (`general.json`) — shared operational knowledge (tool usage, device info, routines)
+
+Memories are injected into the system prompt by the bridge with budget truncation (personal: 4000 chars, general: 3000 chars, newest first).
+
+### TTS Voice Routing
+
+The bridge writes the active character's TTS config to `~/homeai-data/active-tts-voice.json` before each request. The Wyoming TTS server reads this state file to determine which engine/voice to use:
+- **Kokoro** — local, fast, uses `kokoro_voice` field (e.g., `af_heart`)
+- **ElevenLabs** — cloud, uses `elevenlabs_voice_id` + `elevenlabs_model`, returns PCM 24kHz
+
+This works for both ESP32/HA pipeline and dashboard chat.
 
 ---
 
 ## Project Priorities
 
-1. **Foundation** — Docker stack up (Home Assistant, Open WebUI, Portainer, Uptime Kuma)
+1. **Foundation** — Docker stack up (Home Assistant, Open WebUI, Portainer, Uptime Kuma) ✅
-2. **LLM** — Ollama running with target models, Open WebUI connected
+2. **LLM** — Ollama running with target models, Open WebUI connected ✅
-3. **Voice pipeline** — Whisper → Ollama → Kokoro → Wyoming → Home Assistant
+3. **Voice pipeline** — Whisper → Ollama → Kokoro → Wyoming → Home Assistant ✅
-4. **OpenClaw** — installed, onboarded, connected to Ollama and Home Assistant
+4. **OpenClaw** — installed, onboarded, connected to Ollama and Home Assistant ✅
-5. **ESP32-S3-BOX-3** — ESPHome flash, Wyoming Satellite, LVGL face
+5. **ESP32-S3-BOX-3** — ESPHome flash, Wyoming Satellite, display faces ✅
-6. **Character system** — system prompt wired up, mem0 integrated, voice cloned
+6. **Character system** — schema v2, dashboard editor, memory system, per-character TTS routing ✅
-7. **VTube Studio** — model loaded, WebSocket API bridge written as OpenClaw skill
+7. **Animated visual** — PNG/GIF character visual for the web assistant (initial visual layer)
-8. **ComfyUI** — image generation online, character-consistent model workflows
+8. **Android app** — companion app for mobile access to the assistant
-9. **Extended integrations** — n8n workflows, Music Assistant, Snapcast, Gitea, code-server
+9. **ComfyUI** — image generation online, character-consistent model workflows
-10. **Polish** — Authelia, Tailscale hardening, mobile companion, iOS widgets
+10. **Extended integrations** — n8n workflows, Music Assistant, Snapcast, Gitea, code-server
+11. **Polish** — Authelia, Tailscale hardening, iOS widgets
+
+### Stretch Goals
+- **Live2D / VTube Studio** — full Live2D model with WebSocket API bridge (requires learning Live2D tooling)
 
 ---
 
@@ -133,7 +169,11 @@ Character config JSON (exported from Character Manager) is the single source of
 
 - All Docker compose files: `~/server/docker/`
 - OpenClaw skills: `~/.openclaw/skills/`
-- Character configs: `~/.openclaw/characters/`
+- Character configs: `~/homeai-data/characters/`
+- Character memories: `~/homeai-data/memories/`
+- Conversation history: `~/homeai-data/conversations/`
+- Active TTS state: `~/homeai-data/active-tts-voice.json`
+- Satellite → character map: `~/homeai-data/satellite-map.json`
 - Whisper models: `~/models/whisper/`
 - Ollama models: managed by Ollama at `~/.ollama/models/`
 - ComfyUI models: `~/ComfyUI/models/`

TODO.md

@@ -26,7 +26,7 @@
 - [x] Register local GGUF models via Modelfiles (no download): llama3.3:70b, qwen3:32b, codestral:22b, qwen2.5:7b
 - [x] Register additional models: EVA-LLaMA-3.33-70B, Midnight-Miqu-70B, QwQ-32B, Qwen3.5-35B, Qwen3-Coder-30B, Qwen3-VL-30B, GLM-4.6V-Flash, DeepSeek-R1-8B, gemma-3-27b
 - [x] Add qwen3.5:35b-a3b (MoE, Q8_0) — 26.7 tok/s, recommended for voice pipeline
-- [x] Write model preload script + launchd service (keeps voice model in VRAM permanently)
+- [x] Write model keep-warm daemon + launchd service (pins qwen2.5:7b + $HOMEAI_MEDIUM_MODEL in VRAM, checks every 5 min)
 - [x] Deploy Open WebUI via Docker compose (port 3030)
 - [x] Verify Open WebUI connected to Ollama, all models available
 - [x] Run pipeline benchmark (homeai-voice/scripts/benchmark_pipeline.py) — STT/LLM/TTS latency profiled
@@ -82,7 +82,7 @@
 - [x] Verify full voice → agent → HA action flow
 - [x] Add OpenClaw to Uptime Kuma monitors (Manual user action required)
 
-### P5 · homeai-character *(can start alongside P4)*
+### P5 · homeai-dashboard *(character system + dashboard)*
 
 - [x] Define and write `schema/character.schema.json` (v1)
 - [x] Write `characters/aria.json` — default character
@@ -100,6 +100,15 @@
 - [x] Add character profile management to dashboard — store/switch character configs with attached profile images
 - [x] Add TTS voice preview in character editor — Kokoro preview via OpenClaw bridge with loading state, custom text, stop control
 - [x] Merge homeai-character + homeai-desktop into unified homeai-dashboard (services, chat, characters, editor)
+- [x] Upgrade character schema to v2 — background, dialogue_style, appearance, skills, gaze_presets (auto-migrate v1)
+- [x] Add LLM-assisted character creation via Character MCP server (Fandom/Wikipedia lookup)
+- [x] Add character memory system — personal (per-character) + general (shared) memories with dashboard UI
+- [x] Add conversation history with per-conversation persistence
+- [x] Wire character_id through full pipeline (dashboard → bridge → LLM system prompt)
+- [x] Add TTS text cleaning — strip tags, asterisks, emojis, markdown before synthesis
+- [x] Add per-character TTS voice routing — bridge writes state file, Wyoming server reads it
+- [x] Add ElevenLabs TTS support in Wyoming server — cloud voice synthesis via state file routing
+- [x] Dashboard auto-selects character's TTS engine/voice (Kokoro or ElevenLabs)
 - [ ] Deploy dashboard as Docker container or static site on Mac Mini
 
 ---
@@ -123,50 +132,71 @@
 - [ ] Flash remaining units (bedroom, kitchen)
 - [ ] Document MAC address → room name mapping
 
+### P6b · homeai-rpi (Kitchen Satellite)
+
+- [x] Set up Wyoming Satellite on Raspberry Pi 5 (SELBINA) with ReSpeaker 2-Mics pHAT
+- [x] Write setup.sh — full Pi provisioning (venv, drivers, systemd, scripts)
+- [x] Write deploy.sh — remote deploy/manage from Mac Mini (push-wrapper, test-logs, etc.)
+- [x] Write satellite_wrapper.py — monkey-patches fixing TTS echo, writer race, streaming timeout
+- [x] Test multi-command voice loop without freezing
+
 ---
 
 ## Phase 5 — Visual Layer
 
 ### P7 · homeai-visual
 
-- [ ] Install VTube Studio (Mac App Store)
+#### VTube Studio Expression Bridge
-- [ ] Enable WebSocket API on port 8001
+- [x] Write `vtube-bridge.py` — persistent WebSocket ↔ HTTP bridge daemon (port 8002)
-- [ ] Source/purchase a Live2D model (nizima.com or booth.pm)
+- [x] Write `vtube-ctl` CLI wrapper + OpenClaw skill (`~/.openclaw/skills/vtube-studio/`)
-- [ ] Load model in VTube Studio
+- [x] Wire expression triggers into `openclaw-http-bridge.py` (thinking → idle, speaking → idle)
-- [ ] Create hotkeys for all 8 expression states
+- [x] Add amplitude-based lip sync to `wyoming_kokoro_server.py` (RMS → MouthOpen parameter)
-- [ ] Write `skills/vtube_studio` SKILL.md + implementation
+- [x] Write `test-expressions.py` — auth flow, expression cycle, lip sync sweep, latency test
-- [ ] Run auth flow — click Allow in VTube Studio, save token
+- [x] Write launchd plist + setup.sh for venv creation and service registration
-- [ ] Test all 8 expressions via test script
+- [ ] Install VTube Studio from Mac App Store, enable WebSocket API (port 8001)
-- [ ] Update `aria.json` with real VTube Studio hotkey IDs
+- [ ] Source/purchase Live2D model, load in VTube Studio
-- [ ] Write `lipsync.py` amplitude-based helper
+- [ ] Create 8 expression hotkeys, record UUIDs
-- [ ] Integrate lip sync into OpenClaw TTS dispatch
+- [ ] Run `setup.sh` to create venv, install websockets, load launchd service
-- [ ] Test full pipeline: voice → thinking expression → speaking with lip sync
+- [ ] Run `vtube-ctl auth` — click Allow in VTube Studio
+- [ ] Update `aria.json` with real hotkey UUIDs (replace placeholders)
+- [ ] Run `test-expressions.py --all` — verify expressions + lip sync + latency
 - [ ] Set up VTube Studio mobile (iPhone/iPad) on Tailnet
 
+#### Web Visuals (Dashboard)
+- [ ] Design PNG/GIF character visuals for web assistant (idle, thinking, speaking, etc.)
+- [ ] Integrate animated visuals into homeai-dashboard chat view
+- [ ] Sync visual state to voice pipeline events (listening, processing, responding)
+- [ ] Add expression transitions and idle animations
+
+### P8 · homeai-android
+
+- [ ] Build Android companion app for mobile assistant access
+- [ ] Integrate with OpenClaw bridge API (chat, TTS, STT)
+- [ ] Add character visual display
+- [ ] Push notification support via ntfy/FCM
+
 ---
 
 ## Phase 6 — Image Generation
 
-### P8 · homeai-images
+### P9 · homeai-images (ComfyUI)
 
 - [ ] Clone ComfyUI to `~/ComfyUI/`, install deps in venv
 - [ ] Verify MPS is detected at launch
 - [ ] Write and load launchd plist (`com.homeai.comfyui.plist`)
-- [ ] Download SDXL base model
+- [ ] Download SDXL base model + Flux.1-schnell + ControlNet models
-- [ ] Download Flux.1-schnell
-- [ ] Download ControlNet models (canny, depth)
 - [ ] Test generation via ComfyUI web UI (port 8188)
-- [ ] Build and export `quick.json`, `portrait.json`, `scene.json`, `upscale.json` workflows
+- [ ] Build and export workflow JSONs (quick, portrait, scene, upscale)
 - [ ] Write `skills/comfyui` SKILL.md + implementation
-- [ ] Test skill: "Generate a portrait of Aria looking happy"
 - [ ] Collect character reference images for LoRA training
-- [ ] Train SDXL LoRA with kohya_ss, verify character consistency
 - [ ] Add ComfyUI to Uptime Kuma monitors
 
 ---
 
 ## Phase 7 — Extended Integrations & Polish
 
+### P10 · Integrations & Polish
+
 - [ ] Deploy Music Assistant (Docker), integrate with Home Assistant
 - [ ] Write `skills/music` SKILL.md for OpenClaw
 - [ ] Deploy Snapcast server on Mac Mini
@@ -183,10 +213,24 @@
 
 ---
 
+## Stretch Goals
+
+### Live2D / VTube Studio
+
+- [ ] Learn Live2D modelling toolchain (Live2D Cubism Editor)
+- [ ] Install VTube Studio (Mac App Store), enable WebSocket API on port 8001
+- [ ] Source/commission a Live2D model (nizima.com or booth.pm)
+- [ ] Create hotkeys for expression states
+- [ ] Write `skills/vtube_studio` SKILL.md + implementation
+- [ ] Write `lipsync.py` amplitude-based helper
+- [ ] Integrate lip sync into OpenClaw TTS dispatch
+- [ ] Set up VTube Studio mobile (iPhone/iPad) on Tailnet
+
+---
+
 ## Open Decisions
 
 - [ ] Confirm character name (determines wake word training)
-- [ ] Live2D model: purchase off-the-shelf or commission custom?
 - [ ] mem0 backend: Chroma (simple) vs Qdrant Docker (better semantic search)?
 - [ ] Snapcast output: ESP32 built-in speakers or dedicated audio hardware per room?
 - [ ] Authelia user store: local file vs LDAP?

homeai-agent/custom_components/openclaw_conversation/const.py

@@ -12,7 +12,7 @@ CONF_TIMEOUT = "timeout"
 DEFAULT_HOST = "10.0.0.101"
 DEFAULT_PORT = 8081  # OpenClaw HTTP Bridge (not 8080 gateway)
 DEFAULT_AGENT = "main"
-DEFAULT_TIMEOUT = 120
+DEFAULT_TIMEOUT = 200  # Must exceed bridge cold timeout (180s)
 
 # API endpoints
 OPENCLAW_API_PATH = "/api/agent/message"

homeai-agent/custom_components/openclaw_conversation/conversation.py

@@ -77,7 +77,11 @@ class OpenClawAgent(AbstractConversationAgent):
         _LOGGER.debug("Processing message: %s", text)
         
         try:
-            response_text = await self._call_openclaw(text)
+            response_text = await self._call_openclaw(
+                text,
+                satellite_id=getattr(user_input, "satellite_id", None),
+                device_id=getattr(user_input, "device_id", None),
+            )
 
             # Create proper IntentResponse for Home Assistant
             intent_response = IntentResponse(language=user_input.language or "en")
@@ -96,13 +100,14 @@ class OpenClawAgent(AbstractConversationAgent):
                 conversation_id=conversation_id,
             )
 
-    async def _call_openclaw(self, message: str) -> str:
+    async def _call_openclaw(self, message: str, satellite_id: str = None, device_id: str = None) -> str:
         """Call OpenClaw API and return the response."""
         url = f"http://{self.host}:{self.port}{OPENCLAW_API_PATH}"
 
         payload = {
             "message": message,
             "agent": self.agent_name,
+            "satellite_id": satellite_id or device_id,
         }
         
         session = async_get_clientsession(self.hass)

homeai-agent/launchd/com.homeai.openclaw-bridge.plist

@@ -35,6 +35,8 @@
   <dict>
     <key>PATH</key>
     <string>/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin</string>
+    <key>ELEVENLABS_API_KEY</key>
+    <string>sk_ec10e261c6190307a37aa161a9583504dcf25a0cabe5dbd5</string>
   </dict>
 </dict>
 </plist>

homeai-agent/launchd/com.homeai.openclaw.plist

@@ -28,6 +28,8 @@
     <string>eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJmZGQ1NzZlYWNkMTU0ZTY2ODY1OTkzYTlhNTIxM2FmNyIsImlhdCI6MTc3MjU4ODYyOCwiZXhwIjoyMDg3OTQ4NjI4fQ.CTAU1EZgpVLp_aRnk4vg6cQqwS5N-p8jQkAAXTxFmLY</string>
     <key>HASS_TOKEN</key>
     <string>eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJmZGQ1NzZlYWNkMTU0ZTY2ODY1OTkzYTlhNTIxM2FmNyIsImlhdCI6MTc3MjU4ODYyOCwiZXhwIjoyMDg3OTQ4NjI4fQ.CTAU1EZgpVLp_aRnk4vg6cQqwS5N-p8jQkAAXTxFmLY</string>
+    <key>GAZE_API_KEY</key>
+    <string>e63401f17e4845e1059f830267f839fe7fc7b6083b1cb1730863318754d799f4</string>
   </dict>
 
   <key>RunAtLoad</key>

homeai-agent/openclaw-http-bridge.py

@@ -24,9 +24,12 @@ Endpoints:
 
 import argparse
 import json
+import os
 import subprocess
 import sys
 import asyncio
+import urllib.request
+import threading
 from http.server import HTTPServer, BaseHTTPRequestHandler
 from socketserver import ThreadingMixIn
 from urllib.parse import urlparse
@@ -40,19 +43,222 @@ from wyoming.asr import Transcribe, Transcript
 from wyoming.audio import AudioStart, AudioChunk, AudioStop
 from wyoming.info import Info
 
+# Timeout settings (seconds)
+TIMEOUT_WARM = 120   # Model already loaded in VRAM
+TIMEOUT_COLD = 180   # Model needs loading first (~10-20s load + inference)
+OLLAMA_PS_URL = "http://localhost:11434/api/ps"
+VTUBE_BRIDGE_URL = "http://localhost:8002"
 
-def load_character_prompt() -> str:
+
-    """Load the active character system prompt."""
+def _vtube_fire_and_forget(path: str, data: dict):
-    character_path = Path.home() / ".openclaw" / "characters" / "aria.json"
+    """Send a non-blocking POST to the VTube Studio bridge. Failures are silent."""
+    def _post():
+        try:
+            body = json.dumps(data).encode()
+            req = urllib.request.Request(
+                f"{VTUBE_BRIDGE_URL}{path}",
+                data=body,
+                headers={"Content-Type": "application/json"},
+                method="POST",
+            )
+            urllib.request.urlopen(req, timeout=2)
+        except Exception:
+            pass  # bridge may not be running — that's fine
+    threading.Thread(target=_post, daemon=True).start()
+
+
+def is_model_warm() -> bool:
+    """Check if the default Ollama model is already loaded in VRAM."""
+    try:
+        req = urllib.request.Request(OLLAMA_PS_URL)
+        with urllib.request.urlopen(req, timeout=2) as resp:
+            data = json.loads(resp.read())
+            return len(data.get("models", [])) > 0
+    except Exception:
+        # If we can't reach Ollama, assume cold (safer longer timeout)
+        return False
+
+
+CHARACTERS_DIR = Path("/Users/aodhan/homeai-data/characters")
+SATELLITE_MAP_PATH = Path("/Users/aodhan/homeai-data/satellite-map.json")
+MEMORIES_DIR = Path("/Users/aodhan/homeai-data/memories")
+ACTIVE_TTS_VOICE_PATH = Path("/Users/aodhan/homeai-data/active-tts-voice.json")
+
+
+def clean_text_for_tts(text: str) -> str:
+    """Strip content that shouldn't be spoken: tags, asterisks, emojis, markdown."""
+    # Remove HTML/XML tags and their content for common non-spoken tags
+    text = re.sub(r'<[^>]+>', '', text)
+    # Remove content between asterisks (actions/emphasis markup like *sighs*)
+    text = re.sub(r'\*[^*]+\*', '', text)
+    # Remove markdown bold/italic markers that might remain
+    text = re.sub(r'[*_]{1,3}', '', text)
+    # Remove markdown headers
+    text = re.sub(r'^#{1,6}\s+', '', text, flags=re.MULTILINE)
+    # Remove markdown links [text](url) → keep text
+    text = re.sub(r'\[([^\]]+)\]\([^)]+\)', r'\1', text)
+    # Remove bare URLs
+    text = re.sub(r'https?://\S+', '', text)
+    # Remove code blocks and inline code
+    text = re.sub(r'```[\s\S]*?```', '', text)
+    text = re.sub(r'`[^`]+`', '', text)
+    # Remove emojis
+    text = re.sub(
+        r'[\U0001F600-\U0001F64F\U0001F300-\U0001F5FF\U0001F680-\U0001F6FF'
+        r'\U0001F1E0-\U0001F1FF\U0001F900-\U0001F9FF\U0001FA00-\U0001FAFF'
+        r'\U00002702-\U000027B0\U0000FE00-\U0000FE0F\U0000200D'
+        r'\U00002600-\U000026FF\U00002300-\U000023FF]+', '', text
+    )
+    # Collapse multiple spaces/newlines
+    text = re.sub(r'\n{2,}', '\n', text)
+    text = re.sub(r'[ \t]{2,}', ' ', text)
+    return text.strip()
+
+
+def load_satellite_map() -> dict:
+    """Load the satellite-to-character mapping."""
+    try:
+        with open(SATELLITE_MAP_PATH) as f:
+            return json.load(f)
+    except Exception:
+        return {"default": "aria_default", "satellites": {}}
+
+
+def set_active_tts_voice(character_id: str, tts_config: dict):
+    """Write the active TTS config to a state file for the Wyoming TTS server to read."""
+    try:
+        ACTIVE_TTS_VOICE_PATH.parent.mkdir(parents=True, exist_ok=True)
+        state = {
+            "character_id": character_id,
+            "engine": tts_config.get("engine", "kokoro"),
+            "kokoro_voice": tts_config.get("kokoro_voice", ""),
+            "elevenlabs_voice_id": tts_config.get("elevenlabs_voice_id", ""),
+            "elevenlabs_model": tts_config.get("elevenlabs_model", "eleven_multilingual_v2"),
+            "speed": tts_config.get("speed", 1),
+        }
+        with open(ACTIVE_TTS_VOICE_PATH, "w") as f:
+            json.dump(state, f)
+    except Exception as e:
+        print(f"[OpenClaw Bridge] Warning: could not write active TTS config: {e}")
+
+
+def resolve_character_id(satellite_id: str = None) -> str:
+    """Resolve a satellite ID to a character profile ID."""
+    sat_map = load_satellite_map()
+    if satellite_id and satellite_id in sat_map.get("satellites", {}):
+        return sat_map["satellites"][satellite_id]
+    return sat_map.get("default", "aria_default")
+
+
+def load_character(character_id: str = None) -> dict:
+    """Load a character profile by ID. Returns the full character data dict."""
+    if not character_id:
+        character_id = resolve_character_id()
+    safe_id = character_id.replace("/", "_")
+    character_path = CHARACTERS_DIR / f"{safe_id}.json"
     if not character_path.exists():
-        return ""
+        return {}
     try:
         with open(character_path) as f:
-            data = json.load(f)
+            profile = json.load(f)
-            return data.get("system_prompt", "")
+            return profile.get("data", {})
     except Exception:
+        return {}
+
+
+def load_character_prompt(satellite_id: str = None, character_id: str = None) -> str:
+    """Load the full system prompt for a character, resolved by satellite or explicit ID.
+    Builds a rich prompt from system_prompt + profile fields (background, dialogue_style, etc.)."""
+    if not character_id:
+        character_id = resolve_character_id(satellite_id)
+    char = load_character(character_id)
+    if not char:
         return ""
 
+    sections = []
+
+    # Core system prompt
+    prompt = char.get("system_prompt", "")
+    if prompt:
+        sections.append(prompt)
+
+    # Character profile fields
+    profile_parts = []
+    if char.get("background"):
+        profile_parts.append(f"## Background\n{char['background']}")
+    if char.get("appearance"):
+        profile_parts.append(f"## Appearance\n{char['appearance']}")
+    if char.get("dialogue_style"):
+        profile_parts.append(f"## Dialogue Style\n{char['dialogue_style']}")
+    if char.get("skills"):
+        skills = char["skills"]
+        if isinstance(skills, list):
+            skills_text = ", ".join(skills[:15])
+        else:
+            skills_text = str(skills)
+        profile_parts.append(f"## Skills & Interests\n{skills_text}")
+    if profile_parts:
+        sections.append("[Character Profile]\n" + "\n\n".join(profile_parts))
+
+    # Character metadata
+    meta_lines = []
+    if char.get("display_name"):
+        meta_lines.append(f"Your name is: {char['display_name']}")
+    # Support both v1 (gaze_preset string) and v2 (gaze_presets array)
+    gaze_presets = char.get("gaze_presets", [])
+    if gaze_presets and isinstance(gaze_presets, list):
+        for gp in gaze_presets:
+            preset = gp.get("preset", "")
+            trigger = gp.get("trigger", "self-portrait")
+            if preset:
+                meta_lines.append(f"GAZE preset '{preset}' — use for: {trigger}")
+    elif char.get("gaze_preset"):
+        meta_lines.append(f"Your gaze_preset for self-portraits is: {char['gaze_preset']}")
+    if meta_lines:
+        sections.append("[Character Metadata]\n" + "\n".join(meta_lines))
+
+    # Memories (personal + general)
+    personal, general = load_memories(character_id)
+    if personal:
+        sections.append("[Personal Memories]\n" + "\n".join(f"- {m}" for m in personal))
+    if general:
+        sections.append("[General Knowledge]\n" + "\n".join(f"- {m}" for m in general))
+
+    return "\n\n".join(sections)
+
+
+def load_memories(character_id: str) -> tuple[list[str], list[str]]:
+    """Load personal (per-character) and general memories.
+    Returns (personal_contents, general_contents) truncated to fit context budget."""
+    PERSONAL_BUDGET = 4000  # max chars for personal memories in prompt
+    GENERAL_BUDGET = 3000   # max chars for general memories in prompt
+
+    def _read_memories(path: Path, budget: int) -> list[str]:
+        try:
+            with open(path) as f:
+                data = json.load(f)
+        except Exception:
+            return []
+        memories = data.get("memories", [])
+        # Sort newest first
+        memories.sort(key=lambda m: m.get("createdAt", ""), reverse=True)
+        result = []
+        used = 0
+        for m in memories:
+            content = m.get("content", "").strip()
+            if not content:
+                continue
+            if used + len(content) > budget:
+                break
+            result.append(content)
+            used += len(content)
+        return result
+
+    safe_id = character_id.replace("/", "_")
+    personal = _read_memories(MEMORIES_DIR / "personal" / f"{safe_id}.json", PERSONAL_BUDGET)
+    general = _read_memories(MEMORIES_DIR / "general.json", GENERAL_BUDGET)
+    return personal, general
+
 
 class OpenClawBridgeHandler(BaseHTTPRequestHandler):
     """HTTP request handler for OpenClaw bridge."""
@@ -95,7 +301,7 @@ class OpenClawBridgeHandler(BaseHTTPRequestHandler):
         self._send_json_response(404, {"error": "Not found"})
 
     def _handle_tts_request(self):
-        """Handle TTS request and return wav audio."""
+        """Handle TTS request and return audio. Routes to Kokoro or ElevenLabs based on engine."""
         content_length = int(self.headers.get("Content-Length", 0))
         if content_length == 0:
             self._send_json_response(400, {"error": "Empty body"})
@@ -109,30 +315,64 @@ class OpenClawBridgeHandler(BaseHTTPRequestHandler):
             return
 
         text = data.get("text", "Hello, this is a test.")
-        # Strip emojis so TTS doesn't try to read them out
+        text = clean_text_for_tts(text)
-        text = re.sub(
-            r'[\U0001F600-\U0001F64F\U0001F300-\U0001F5FF\U0001F680-\U0001F6FF'
-            r'\U0001F1E0-\U0001F1FF\U0001F900-\U0001F9FF\U0001FA00-\U0001FAFF'
-            r'\U00002702-\U000027B0\U0000FE00-\U0000FE0F\U0000200D'
-            r'\U00002600-\U000026FF\U00002300-\U000023FF]+', '', text
-        ).strip()
         voice = data.get("voice", "af_heart")
+        engine = data.get("engine", "kokoro")
 
         try:
-            # Run the async Wyoming client
+            # Signal avatar: speaking
-            audio_bytes = asyncio.run(self._synthesize_audio(text, voice))
+            _vtube_fire_and_forget("/expression", {"event": "speaking"})
+
+            if engine == "elevenlabs":
+                audio_bytes, content_type = self._synthesize_elevenlabs(text, voice, data.get("model"))
+            else:
+                # Default: local Kokoro via Wyoming
+                audio_bytes = asyncio.run(self._synthesize_audio(text, voice))
+                content_type = "audio/wav"
+
+            # Signal avatar: idle
+            _vtube_fire_and_forget("/expression", {"event": "idle"})
 
-            # Send WAV response
             self.send_response(200)
-            self.send_header("Content-Type", "audio/wav")
+            self.send_header("Content-Type", content_type)
-            # Allow CORS for local testing from Vite
             self.send_header("Access-Control-Allow-Origin", "*")
             self.end_headers()
             self.wfile.write(audio_bytes)
 
         except Exception as e:
+            _vtube_fire_and_forget("/expression", {"event": "error"})
             self._send_json_response(500, {"error": str(e)})
 
+    def _synthesize_elevenlabs(self, text: str, voice_id: str, model: str = None) -> tuple[bytes, str]:
+        """Call ElevenLabs TTS API and return (audio_bytes, content_type)."""
+        api_key = os.environ.get("ELEVENLABS_API_KEY", "")
+        if not api_key:
+            raise RuntimeError("ELEVENLABS_API_KEY not set in environment")
+        if not voice_id:
+            raise RuntimeError("No ElevenLabs voice ID provided")
+
+        model = model or "eleven_multilingual_v2"
+        url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}"
+        payload = json.dumps({
+            "text": text,
+            "model_id": model,
+            "voice_settings": {"stability": 0.5, "similarity_boost": 0.75},
+        }).encode()
+
+        req = urllib.request.Request(
+            url,
+            data=payload,
+            headers={
+                "Content-Type": "application/json",
+                "xi-api-key": api_key,
+                "Accept": "audio/mpeg",
+            },
+            method="POST",
+        )
+        with urllib.request.urlopen(req, timeout=30) as resp:
+            audio_bytes = resp.read()
+        return audio_bytes, "audio/mpeg"
+
     def do_OPTIONS(self):
         """Handle CORS preflight requests."""
         self.send_response(204)
@@ -264,6 +504,43 @@ class OpenClawBridgeHandler(BaseHTTPRequestHandler):
         print(f"[OpenClaw Bridge] Wake word detected: {wake_word_data.get('wake_word', 'unknown')}")
         self._send_json_response(200, {"status": "ok", "message": "Wake word received"})
 
+    @staticmethod
+    def _call_openclaw(message: str, agent: str, timeout: int) -> str:
+        """Call OpenClaw CLI and return stdout."""
+        result = subprocess.run(
+            ["/opt/homebrew/bin/openclaw", "agent", "--message", message, "--agent", agent],
+            capture_output=True,
+            text=True,
+            timeout=timeout,
+            check=True,
+        )
+        return result.stdout.strip()
+
+    @staticmethod
+    def _needs_followup(response: str) -> bool:
+        """Detect if the model promised to act but didn't actually do it.
+        Returns True if the response looks like a 'will do' without a result."""
+        if not response:
+            return False
+        resp_lower = response.lower()
+        # If the response contains a URL or JSON-like output, it probably completed
+        if "http://" in response or "https://" in response or '"status"' in response:
+            return False
+        # If it contains a tool result indicator (ha-ctl output, gaze-ctl output)
+        if any(kw in resp_lower for kw in ["image_url", "seed", "entity_id", "state:", "turned on", "turned off"]):
+            return False
+        # Detect promise-like language without substance
+        promise_phrases = [
+            "let me", "i'll ", "i will ", "sure thing", "sure,", "right away",
+            "generating", "one moment", "working on", "hang on", "just a moment",
+            "on it", "let me generate", "let me create",
+        ]
+        has_promise = any(phrase in resp_lower for phrase in promise_phrases)
+        # Short responses with promise language are likely incomplete
+        if has_promise and len(response) < 200:
+            return True
+        return False
+
     def _handle_agent_request(self):
         """Handle agent message request."""
         content_length = int(self.headers.get("Content-Length", 0))
@@ -280,29 +557,63 @@ class OpenClawBridgeHandler(BaseHTTPRequestHandler):
 
         message = data.get("message")
         agent = data.get("agent", "main")
+        satellite_id = data.get("satellite_id")
+        explicit_character_id = data.get("character_id")
 
         if not message:
             self._send_json_response(400, {"error": "Message is required"})
             return
 
-        # Inject system prompt
+        # Resolve character: explicit ID > satellite mapping > default
-        system_prompt = load_character_prompt()
+        if explicit_character_id:
+            character_id = explicit_character_id
+        else:
+            character_id = resolve_character_id(satellite_id)
+        system_prompt = load_character_prompt(character_id=character_id)
+
+        # Set the active TTS config for the Wyoming server to pick up
+        char = load_character(character_id)
+        tts_config = char.get("tts", {})
+        if tts_config:
+            set_active_tts_voice(character_id, tts_config)
+            engine = tts_config.get("engine", "kokoro")
+            voice_label = tts_config.get("kokoro_voice", "") if engine == "kokoro" else tts_config.get("elevenlabs_voice_id", "")
+            print(f"[OpenClaw Bridge] Active TTS: {engine} / {voice_label}")
+
+        if satellite_id:
+            print(f"[OpenClaw Bridge] Satellite: {satellite_id} → character: {character_id}")
+        elif explicit_character_id:
+            print(f"[OpenClaw Bridge] Character: {character_id}")
         if system_prompt:
             message = f"System Context: {system_prompt}\n\nUser Request: {message}"
 
+        # Check if model is warm to set appropriate timeout
+        warm = is_model_warm()
+        timeout = TIMEOUT_WARM if warm else TIMEOUT_COLD
+        print(f"[OpenClaw Bridge] Model {'warm' if warm else 'cold'}, timeout={timeout}s")
+
+        # Signal avatar: thinking
+        _vtube_fire_and_forget("/expression", {"event": "thinking"})
+
         # Call OpenClaw CLI (use full path for launchd compatibility)
         try:
-            result = subprocess.run(
+            response_text = self._call_openclaw(message, agent, timeout)
-                ["/opt/homebrew/bin/openclaw", "agent", "--message", message, "--agent", agent],
+
-                capture_output=True,
+            # Re-prompt if the model promised to act but didn't call a tool.
-                text=True,
+            # Detect "I'll do X" / "Let me X" responses that lack any result.
-                timeout=120,
+            if self._needs_followup(response_text):
-                check=True
+                print(f"[OpenClaw Bridge] Response looks like a promise without action, re-prompting")
-            )
+                followup = (
-            response_text = result.stdout.strip()
+                    "You just said you would do something but didn't actually call the exec tool. "
+                    "Do NOT explain what you will do — call the tool NOW using exec and return the result."
+                )
+                response_text = self._call_openclaw(followup, agent, timeout)
+
+            # Signal avatar: idle (TTS handler will override to 'speaking' if voice is used)
+            _vtube_fire_and_forget("/expression", {"event": "idle"})
             self._send_json_response(200, {"response": response_text})
         except subprocess.TimeoutExpired:
-            self._send_json_response(504, {"error": "OpenClaw command timed out"})
+            self._send_json_response(504, {"error": f"OpenClaw command timed out after {timeout}s (model was {'warm' if warm else 'cold'})"})
         except subprocess.CalledProcessError as e:
             error_msg = e.stderr.strip() if e.stderr else "OpenClaw command failed"
             self._send_json_response(500, {"error": error_msg})

homeai-dashboard/launchd/com.homeai.dashboard.plist

@@ -24,6 +24,8 @@
     <string>/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin</string>
     <key>HOME</key>
     <string>/Users/aodhan</string>
+    <key>GAZE_API_KEY</key>
+    <string>e63401f17e4845e1059f830267f839fe7fc7b6083b1cb1730863318754d799f4</string>
   </dict>
 
   <key>RunAtLoad</key>

homeai-dashboard/schema/character.schema.json

@@ -1,15 +1,24 @@
 {
   "$schema": "http://json-schema.org/draft-07/schema#",
   "title": "HomeAI Character Config",
-  "version": "1",
+  "version": "2",
   "type": "object",
   "required": ["schema_version", "name", "system_prompt", "tts"],
   "properties": {
-    "schema_version": { "type": "integer", "const": 1 },
+    "schema_version": { "type": "integer", "enum": [1, 2] },
     "name": { "type": "string" },
     "display_name": { "type": "string" },
     "description": { "type": "string" },
 
+    "background": { "type": "string", "description": "Backstory, lore, or general prompt enrichment" },
+    "dialogue_style": { "type": "string", "description": "How the persona speaks or reacts, with example lines" },
+    "appearance": { "type": "string", "description": "Physical description, also used for image prompting" },
+    "skills": {
+      "type": "array",
+      "description": "Topics the persona specialises in or enjoys talking about",
+      "items": { "type": "string" }
+    },
+
     "system_prompt": { "type": "string" },
 
     "model_overrides": {
@@ -31,35 +40,21 @@
         "voice_ref_path": { "type": "string" },
         "kokoro_voice": { "type": "string" },
         "elevenlabs_voice_id": { "type": "string" },
+        "elevenlabs_voice_name": { "type": "string" },
         "elevenlabs_model": { "type": "string", "default": "eleven_monolingual_v1" },
         "speed": { "type": "number", "default": 1.0 }
       }
     },
 
-    "live2d_expressions": {
+    "gaze_presets": {
-      "type": "object",
+      "type": "array",
-      "description": "Maps semantic state to VTube Studio hotkey ID",
+      "description": "GAZE image generation presets with trigger conditions",
-      "properties": {
+      "items": {
-        "idle":      { "type": "string" },
-        "listening": { "type": "string" },
-        "thinking":  { "type": "string" },
-        "speaking":  { "type": "string" },
-        "happy":     { "type": "string" },
-        "sad":       { "type": "string" },
-        "surprised": { "type": "string" },
-        "error":     { "type": "string" }
-      }
-    },
-
-    "vtube_ws_triggers": {
-      "type": "object",
-      "description": "VTube Studio WebSocket actions keyed by event name",
-      "additionalProperties": {
         "type": "object",
+        "required": ["preset"],
         "properties": {
-          "type": { "type": "string", "enum": ["hotkey", "parameter"] },
+          "preset": { "type": "string" },
-          "id":   { "type": "string" },
+          "trigger": { "type": "string", "default": "self-portrait" }
-          "value": { "type": "number" }
         }
       }
     },
@@ -78,5 +73,6 @@
     },
 
     "notes": { "type": "string" }
-  }
+  },
+  "additionalProperties": true
 }

homeai-dashboard/src/App.jsx

@@ -3,6 +3,7 @@ import Dashboard from './pages/Dashboard';
 import Chat from './pages/Chat';
 import Characters from './pages/Characters';
 import Editor from './pages/Editor';
+import Memories from './pages/Memories';
 
 function NavItem({ to, children, icon }) {
   return (
@@ -77,6 +78,17 @@ function Layout({ children }) {
             Characters
           </NavItem>
 
+          <NavItem
+            to="/memories"
+            icon={
+              <svg className="w-5 h-5" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={1.5}>
+                <path strokeLinecap="round" strokeLinejoin="round" d="M12 18v-5.25m0 0a6.01 6.01 0 001.5-.189m-1.5.189a6.01 6.01 0 01-1.5-.189m3.75 7.478a12.06 12.06 0 01-4.5 0m3.75 2.383a14.406 14.406 0 01-3 0M14.25 18v-.192c0-.983.658-1.823 1.508-2.316a7.5 7.5 0 10-7.517 0c.85.493 1.509 1.333 1.509 2.316V18" />
+              </svg>
+            }
+          >
+            Memories
+          </NavItem>
+
           <NavItem
             to="/editor"
             icon={
@@ -113,6 +125,7 @@ function App() {
           <Route path="/" element={<div className="flex-1 overflow-y-auto p-8"><div className="max-w-6xl mx-auto"><Dashboard /></div></div>} />
           <Route path="/chat" element={<Chat />} />
           <Route path="/characters" element={<div className="flex-1 overflow-y-auto p-8"><div className="max-w-6xl mx-auto"><Characters /></div></div>} />
+          <Route path="/memories" element={<div className="flex-1 overflow-y-auto p-8"><div className="max-w-6xl mx-auto"><Memories /></div></div>} />
           <Route path="/editor" element={<div className="flex-1 overflow-y-auto p-8"><div className="max-w-6xl mx-auto"><Editor /></div></div>} />
         </Routes>
       </Layout>

homeai-dashboard/src/components/ChatPanel.jsx

@@ -2,8 +2,10 @@ import { useEffect, useRef } from 'react'
 import MessageBubble from './MessageBubble'
 import ThinkingIndicator from './ThinkingIndicator'
 
-export default function ChatPanel({ messages, isLoading, onReplay }) {
+export default function ChatPanel({ messages, isLoading, onReplay, character }) {
   const bottomRef = useRef(null)
+  const name = character?.name || 'AI'
+  const image = character?.image || null
 
   useEffect(() => {
     bottomRef.current?.scrollIntoView({ behavior: 'smooth' })
@@ -13,10 +15,14 @@ export default function ChatPanel({ messages, isLoading, onReplay }) {
     return (
       <div className="flex-1 flex items-center justify-center">
         <div className="text-center">
-          <div className="w-16 h-16 rounded-full bg-indigo-600/20 flex items-center justify-center mx-auto mb-4">
+          {image ? (
-            <span className="text-indigo-400 text-2xl">AI</span>
+            <img src={image} alt={name} className="w-20 h-20 rounded-full object-cover mx-auto mb-4 ring-2 ring-indigo-500/30" />
-          </div>
+          ) : (
-          <h2 className="text-xl font-medium text-gray-200 mb-2">Hi, I'm Aria</h2>
+            <div className="w-20 h-20 rounded-full bg-indigo-600/20 flex items-center justify-center mx-auto mb-4">
+              <span className="text-indigo-400 text-2xl">{name[0]}</span>
+            </div>
+          )}
+          <h2 className="text-xl font-medium text-gray-200 mb-2">Hi, I'm {name}</h2>
           <p className="text-gray-500 text-sm">Type a message or press the mic to talk</p>
         </div>
       </div>
@@ -26,9 +32,9 @@ export default function ChatPanel({ messages, isLoading, onReplay }) {
   return (
     <div className="flex-1 overflow-y-auto py-4">
       {messages.map((msg) => (
-        <MessageBubble key={msg.id} message={msg} onReplay={onReplay} />
+        <MessageBubble key={msg.id} message={msg} onReplay={onReplay} character={character} />
       ))}
-      {isLoading && <ThinkingIndicator />}
+      {isLoading && <ThinkingIndicator character={character} />}
       <div ref={bottomRef} />
     </div>
   )

homeai-dashboard/src/components/ConversationList.jsx

@@ -0,0 +1,70 @@
+function timeAgo(dateStr) {
+  if (!dateStr) return ''
+  const diff = Date.now() - new Date(dateStr).getTime()
+  const mins = Math.floor(diff / 60000)
+  if (mins < 1) return 'just now'
+  if (mins < 60) return `${mins}m ago`
+  const hours = Math.floor(mins / 60)
+  if (hours < 24) return `${hours}h ago`
+  const days = Math.floor(hours / 24)
+  return `${days}d ago`
+}
+
+export default function ConversationList({ conversations, activeId, onCreate, onSelect, onDelete }) {
+  return (
+    <div className="w-72 border-r border-gray-800 flex flex-col bg-gray-950 shrink-0">
+      {/* New chat button */}
+      <div className="p-3 border-b border-gray-800">
+        <button
+          onClick={onCreate}
+          className="w-full flex items-center justify-center gap-2 px-3 py-2 bg-indigo-600 hover:bg-indigo-500 text-white text-sm rounded-lg transition-colors"
+        >
+          <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+            <path strokeLinecap="round" strokeLinejoin="round" d="M12 4.5v15m7.5-7.5h-15" />
+          </svg>
+          New chat
+        </button>
+      </div>
+
+      {/* Conversation list */}
+      <div className="flex-1 overflow-y-auto">
+        {conversations.length === 0 ? (
+          <p className="text-xs text-gray-600 text-center py-6">No conversations yet</p>
+        ) : (
+          conversations.map(conv => (
+            <div
+              key={conv.id}
+              onClick={() => onSelect(conv.id)}
+              className={`group flex items-start gap-2 px-3 py-2.5 cursor-pointer border-b border-gray-800/50 transition-colors ${
+                conv.id === activeId
+                  ? 'bg-gray-800 text-white'
+                  : 'text-gray-400 hover:bg-gray-800/50 hover:text-gray-200'
+              }`}
+            >
+              <div className="flex-1 min-w-0">
+                <p className="text-sm truncate">
+                  {conv.title || 'New conversation'}
+                </p>
+                <div className="flex items-center gap-2 mt-0.5">
+                  {conv.characterName && (
+                    <span className="text-xs text-indigo-400/70">{conv.characterName}</span>
+                  )}
+                  <span className="text-xs text-gray-600">{timeAgo(conv.updatedAt)}</span>
+                </div>
+              </div>
+              <button
+                onClick={(e) => { e.stopPropagation(); onDelete(conv.id) }}
+                className="opacity-0 group-hover:opacity-100 p-1 text-gray-500 hover:text-red-400 transition-all shrink-0 mt-0.5"
+                title="Delete"
+              >
+                <svg className="w-3.5 h-3.5" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+                  <path strokeLinecap="round" strokeLinejoin="round" d="M14.74 9l-.346 9m-4.788 0L9.26 9m9.968-3.21c.342.052.682.107 1.022.166m-1.022-.165L18.16 19.673a2.25 2.25 0 01-2.244 2.077H8.084a2.25 2.25 0 01-2.244-2.077L4.772 5.79m14.456 0a48.108 48.108 0 00-3.478-.397m-12 .562c.34-.059.68-.114 1.022-.165m0 0a48.11 48.11 0 013.478-.397m7.5 0v-.916c0-1.18-.91-2.164-2.09-2.201a51.964 51.964 0 00-3.32 0c-1.18.037-2.09 1.022-2.09 2.201v.916m7.5 0a48.667 48.667 0 00-7.5 0" />
+                </svg>
+              </button>
+            </div>
+          ))
+        )}
+      </div>
+    </div>
+  )
+}

homeai-dashboard/src/components/MessageBubble.jsx

@@ -1,14 +1,100 @@
-export default function MessageBubble({ message, onReplay }) {
+import { useState } from 'react'
+
+function Avatar({ character }) {
+  const name = character?.name || 'AI'
+  const image = character?.image || null
+
+  if (image) {
+    return <img src={image} alt={name} className="w-8 h-8 rounded-full object-cover shrink-0 mt-0.5 ring-1 ring-gray-700" />
+  }
+
+  return (
+    <div className="w-8 h-8 rounded-full bg-indigo-600/20 flex items-center justify-center shrink-0 mt-0.5">
+      <span className="text-indigo-400 text-sm">{name[0]}</span>
+    </div>
+  )
+}
+
+function ImageOverlay({ src, onClose }) {
+  return (
+    <div
+      className="fixed inset-0 z-50 bg-black/80 flex items-center justify-center cursor-zoom-out"
+      onClick={onClose}
+    >
+      <img
+        src={src}
+        alt="Full size"
+        className="max-w-[90vw] max-h-[90vh] object-contain rounded-lg shadow-2xl"
+        onClick={(e) => e.stopPropagation()}
+      />
+      <button
+        onClick={onClose}
+        className="absolute top-4 right-4 text-white/70 hover:text-white transition-colors p-2"
+      >
+        <svg className="w-6 h-6" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+          <path strokeLinecap="round" strokeLinejoin="round" d="M6 18L18 6M6 6l12 12" />
+        </svg>
+      </button>
+    </div>
+  )
+}
+
+const IMAGE_URL_RE = /(https?:\/\/[^\s]+\.(?:png|jpg|jpeg|gif|webp))/gi
+
+function RichContent({ text }) {
+  const [overlayImage, setOverlayImage] = useState(null)
+  const parts = []
+  let lastIndex = 0
+  let match
+
+  IMAGE_URL_RE.lastIndex = 0
+  while ((match = IMAGE_URL_RE.exec(text)) !== null) {
+    if (match.index > lastIndex) {
+      parts.push({ type: 'text', value: text.slice(lastIndex, match.index) })
+    }
+    parts.push({ type: 'image', value: match[1] })
+    lastIndex = IMAGE_URL_RE.lastIndex
+  }
+  if (lastIndex < text.length) {
+    parts.push({ type: 'text', value: text.slice(lastIndex) })
+  }
+
+  if (parts.length === 1 && parts[0].type === 'text') {
+    return <>{text}</>
+  }
+
+  return (
+    <>
+      {parts.map((part, i) =>
+        part.type === 'image' ? (
+          <button
+            key={i}
+            onClick={() => setOverlayImage(part.value)}
+            className="block my-2 cursor-zoom-in"
+          >
+            <img
+              src={part.value}
+              alt="Generated image"
+              className="rounded-xl max-w-full max-h-80 object-contain"
+              loading="lazy"
+            />
+          </button>
+        ) : (
+          <span key={i}>{part.value}</span>
+        )
+      )}
+      {overlayImage && <ImageOverlay src={overlayImage} onClose={() => setOverlayImage(null)} />}
+    </>
+  )
+}
+
+export default function MessageBubble({ message, onReplay, character }) {
   const isUser = message.role === 'user'
 
   return (
     <div className={`flex ${isUser ? 'justify-end' : 'justify-start'} px-4 py-1.5`}>
       <div className={`flex items-start gap-3 max-w-[80%] ${isUser ? 'flex-row-reverse' : ''}`}>
-        {!isUser && (
+        {!isUser && <Avatar character={character} />}
-          <div className="w-8 h-8 rounded-full bg-indigo-600/20 flex items-center justify-center shrink-0 mt-0.5">
-            <span className="text-indigo-400 text-sm">AI</span>
-          </div>
-        )}
         <div>
           <div
             className={`rounded-2xl px-4 py-2.5 text-sm leading-relaxed whitespace-pre-wrap ${
@@ -19,7 +105,7 @@ export default function MessageBubble({ message, onReplay }) {
                   : 'bg-gray-800 text-gray-100'
             }`}
           >
-            {message.content}
+            {isUser ? message.content : <RichContent text={message.content} />}
           </div>
           {!isUser && !message.isError && onReplay && (
             <button

homeai-dashboard/src/components/SettingsDrawer.jsx

@@ -1,8 +1,10 @@
-import { VOICES } from '../lib/constants'
+import { VOICES, TTS_ENGINES } from '../lib/constants'
 
 export default function SettingsDrawer({ isOpen, onClose, settings, onUpdate }) {
   if (!isOpen) return null
 
+  const isKokoro = !settings.ttsEngine || settings.ttsEngine === 'kokoro'
+
   return (
     <>
       <div className="fixed inset-0 bg-black/50 z-40" onClick={onClose} />
@@ -16,18 +18,48 @@ export default function SettingsDrawer({ isOpen, onClose, settings, onUpdate })
           </button>
         </div>
         <div className="flex-1 overflow-y-auto p-4 space-y-5">
+          {/* TTS Engine */}
+          <div>
+            <label className="block text-xs font-medium text-gray-400 mb-1.5">TTS Engine</label>
+            <select
+              value={settings.ttsEngine || 'kokoro'}
+              onChange={(e) => onUpdate('ttsEngine', e.target.value)}
+              className="w-full bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
+            >
+              {TTS_ENGINES.map((e) => (
+                <option key={e.id} value={e.id}>{e.label}</option>
+              ))}
+            </select>
+          </div>
+
           {/* Voice */}
           <div>
             <label className="block text-xs font-medium text-gray-400 mb-1.5">Voice</label>
-            <select
+            {isKokoro ? (
-              value={settings.voice}
+              <select
-              onChange={(e) => onUpdate('voice', e.target.value)}
+                value={settings.voice}
-              className="w-full bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
+                onChange={(e) => onUpdate('voice', e.target.value)}
-            >
+                className="w-full bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
-              {VOICES.map((v) => (
+              >
-                <option key={v.id} value={v.id}>{v.label}</option>
+                {VOICES.map((v) => (
-              ))}
+                  <option key={v.id} value={v.id}>{v.label}</option>
-            </select>
+                ))}
+              </select>
+            ) : (
+              <div>
+                <input
+                  type="text"
+                  value={settings.voice || ''}
+                  onChange={(e) => onUpdate('voice', e.target.value)}
+                  className="w-full bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
+                  placeholder={settings.ttsEngine === 'elevenlabs' ? 'ElevenLabs voice ID' : 'Voice identifier'}
+                  readOnly
+                />
+                <p className="text-xs text-gray-500 mt-1">
+                  Set via active character profile
+                </p>
+              </div>
+            )}
           </div>
 
           {/* Auto TTS */}

homeai-dashboard/src/components/ThinkingIndicator.jsx

@@ -1,9 +1,16 @@
-export default function ThinkingIndicator() {
+export default function ThinkingIndicator({ character }) {
+  const name = character?.name || 'AI'
+  const image = character?.image || null
+
   return (
     <div className="flex items-start gap-3 px-4 py-3">
-      <div className="w-8 h-8 rounded-full bg-indigo-600/20 flex items-center justify-center shrink-0">
+      {image ? (
-        <span className="text-indigo-400 text-sm">AI</span>
+        <img src={image} alt={name} className="w-8 h-8 rounded-full object-cover shrink-0 ring-1 ring-gray-700" />
-      </div>
+      ) : (
+        <div className="w-8 h-8 rounded-full bg-indigo-600/20 flex items-center justify-center shrink-0">
+          <span className="text-indigo-400 text-sm">{name[0]}</span>
+        </div>
+      )}
       <div className="flex items-center gap-1 pt-2.5">
         <span className="w-2 h-2 rounded-full bg-gray-400 animate-[bounce_1.4s_ease-in-out_infinite]" />
         <span className="w-2 h-2 rounded-full bg-gray-400 animate-[bounce_1.4s_ease-in-out_0.2s_infinite]" />

homeai-dashboard/src/hooks/useActiveCharacter.js

@@ -0,0 +1,28 @@
+import { useState, useEffect } from 'react'
+
+const ACTIVE_KEY = 'homeai_active_character'
+
+export function useActiveCharacter() {
+  const [character, setCharacter] = useState(null)
+
+  useEffect(() => {
+    const activeId = localStorage.getItem(ACTIVE_KEY)
+    if (!activeId) return
+
+    fetch(`/api/characters/${activeId}`)
+      .then(r => r.ok ? r.json() : null)
+      .then(profile => {
+        if (profile) {
+          setCharacter({
+            id: profile.id,
+            name: profile.data.display_name || profile.data.name || 'AI',
+            image: profile.image || null,
+            tts: profile.data.tts || null,
+          })
+        }
+      })
+      .catch(() => {})
+  }, [])
+
+  return character
+}

homeai-dashboard/src/hooks/useChat.js

@@ -1,45 +1,124 @@
-import { useState, useCallback } from 'react'
+import { useState, useCallback, useEffect, useRef } from 'react'
 import { sendMessage } from '../lib/api'
+import { getConversation, saveConversation } from '../lib/conversationApi'
 
-export function useChat() {
+export function useChat(conversationId, conversationMeta, onConversationUpdate) {
   const [messages, setMessages] = useState([])
   const [isLoading, setIsLoading] = useState(false)
+  const [isLoadingConv, setIsLoadingConv] = useState(false)
+  const convRef = useRef(null)
+  const idRef = useRef(conversationId)
 
-  const send = useCallback(async (text) => {
+  // Keep idRef in sync
+  useEffect(() => { idRef.current = conversationId }, [conversationId])
+
+  // Load conversation from server when ID changes
+  useEffect(() => {
+    if (!conversationId) {
+      setMessages([])
+      convRef.current = null
+      return
+    }
+
+    let cancelled = false
+    setIsLoadingConv(true)
+
+    getConversation(conversationId).then(conv => {
+      if (cancelled) return
+      if (conv) {
+        convRef.current = conv
+        setMessages(conv.messages || [])
+      } else {
+        convRef.current = null
+        setMessages([])
+      }
+      setIsLoadingConv(false)
+    }).catch(() => {
+      if (!cancelled) {
+        convRef.current = null
+        setMessages([])
+        setIsLoadingConv(false)
+      }
+    })
+
+    return () => { cancelled = true }
+  }, [conversationId])
+
+  // Persist conversation to server
+  const persist = useCallback(async (updatedMessages, title, overrideId) => {
+    const id = overrideId || idRef.current
+    if (!id) return
+    const now = new Date().toISOString()
+    const conv = {
+      id,
+      title: title || convRef.current?.title || '',
+      characterId: conversationMeta?.characterId || convRef.current?.characterId || '',
+      characterName: conversationMeta?.characterName || convRef.current?.characterName || '',
+      createdAt: convRef.current?.createdAt || now,
+      updatedAt: now,
+      messages: updatedMessages,
+    }
+    convRef.current = conv
+    await saveConversation(conv).catch(() => {})
+    if (onConversationUpdate) {
+      onConversationUpdate(id, {
+        title: conv.title,
+        updatedAt: conv.updatedAt,
+        messageCount: conv.messages.length,
+      })
+    }
+  }, [conversationMeta, onConversationUpdate])
+
+  // send accepts an optional overrideId for when the conversation was just created
+  const send = useCallback(async (text, overrideId) => {
     if (!text.trim() || isLoading) return null
 
-    const userMsg = { id: Date.now(), role: 'user', content: text.trim(), timestamp: new Date() }
+    const userMsg = { id: Date.now(), role: 'user', content: text.trim(), timestamp: new Date().toISOString() }
-    setMessages((prev) => [...prev, userMsg])
+    const isFirstMessage = messages.length === 0
+    const newMessages = [...messages, userMsg]
+    setMessages(newMessages)
     setIsLoading(true)
 
     try {
-      const response = await sendMessage(text.trim())
+      const response = await sendMessage(text.trim(), conversationMeta?.characterId || null)
       const assistantMsg = {
         id: Date.now() + 1,
         role: 'assistant',
         content: response,
-        timestamp: new Date(),
+        timestamp: new Date().toISOString(),
       }
-      setMessages((prev) => [...prev, assistantMsg])
+      const allMessages = [...newMessages, assistantMsg]
+      setMessages(allMessages)
+
+      const title = isFirstMessage
+        ? text.trim().slice(0, 80) + (text.trim().length > 80 ? '...' : '')
+        : undefined
+      await persist(allMessages, title, overrideId)
+
       return response
     } catch (err) {
       const errorMsg = {
         id: Date.now() + 1,
         role: 'assistant',
         content: `Error: ${err.message}`,
-        timestamp: new Date(),
+        timestamp: new Date().toISOString(),
         isError: true,
       }
-      setMessages((prev) => [...prev, errorMsg])
+      const allMessages = [...newMessages, errorMsg]
+      setMessages(allMessages)
+      await persist(allMessages, undefined, overrideId)
       return null
     } finally {
       setIsLoading(false)
     }
-  }, [isLoading])
+  }, [isLoading, messages, persist])
 
-  const clearHistory = useCallback(() => {
+  const clearHistory = useCallback(async () => {
     setMessages([])
-  }, [])
+    if (idRef.current) {
+      await persist([], undefined)
+    }
+  }, [persist])
 
-  return { messages, isLoading, send, clearHistory }
+  return { messages, isLoading, isLoadingConv, send, clearHistory }
 }

homeai-dashboard/src/hooks/useConversations.js

@@ -0,0 +1,66 @@
+import { useState, useEffect, useCallback } from 'react'
+import { listConversations, saveConversation, deleteConversation as deleteConv } from '../lib/conversationApi'
+
+const ACTIVE_KEY = 'homeai_active_conversation'
+
+export function useConversations() {
+  const [conversations, setConversations] = useState([])
+  const [activeId, setActiveId] = useState(() => localStorage.getItem(ACTIVE_KEY) || null)
+  const [isLoading, setIsLoading] = useState(true)
+
+  const loadList = useCallback(async () => {
+    try {
+      const list = await listConversations()
+      setConversations(list)
+    } catch {
+      setConversations([])
+    } finally {
+      setIsLoading(false)
+    }
+  }, [])
+
+  useEffect(() => { loadList() }, [loadList])
+
+  const select = useCallback((id) => {
+    setActiveId(id)
+    if (id) {
+      localStorage.setItem(ACTIVE_KEY, id)
+    } else {
+      localStorage.removeItem(ACTIVE_KEY)
+    }
+  }, [])
+
+  const create = useCallback(async (characterId, characterName) => {
+    const id = `conv_${Date.now()}`
+    const now = new Date().toISOString()
+    const conv = {
+      id,
+      title: '',
+      characterId: characterId || '',
+      characterName: characterName || '',
+      createdAt: now,
+      updatedAt: now,
+      messages: [],
+    }
+    await saveConversation(conv)
+    setConversations(prev => [{ ...conv, messageCount: 0 }, ...prev])
+    select(id)
+    return id
+  }, [select])
+
+  const remove = useCallback(async (id) => {
+    await deleteConv(id)
+    setConversations(prev => prev.filter(c => c.id !== id))
+    if (activeId === id) {
+      select(null)
+    }
+  }, [activeId, select])
+
+  const updateMeta = useCallback((id, updates) => {
+    setConversations(prev => prev.map(c =>
+      c.id === id ? { ...c, ...updates } : c
+    ))
+  }, [])
+
+  return { conversations, activeId, isLoading, select, create, remove, updateMeta, refresh: loadList }
+}

homeai-dashboard/src/hooks/useTtsPlayback.js

@@ -1,7 +1,7 @@
 import { useState, useRef, useCallback } from 'react'
 import { synthesize } from '../lib/api'
 
-export function useTtsPlayback(voice) {
+export function useTtsPlayback(voice, engine = 'kokoro', model = null) {
   const [isPlaying, setIsPlaying] = useState(false)
   const audioCtxRef = useRef(null)
   const sourceRef = useRef(null)
@@ -23,7 +23,7 @@ export function useTtsPlayback(voice) {
 
     setIsPlaying(true)
     try {
-      const audioData = await synthesize(text, voice)
+      const audioData = await synthesize(text, voice, engine, model)
       const ctx = getAudioContext()
       if (ctx.state === 'suspended') await ctx.resume()
 
@@ -42,7 +42,7 @@ export function useTtsPlayback(voice) {
       console.error('TTS playback error:', err)
       setIsPlaying(false)
     }
-  }, [voice])
+  }, [voice, engine, model])
 
   const stop = useCallback(() => {
     if (sourceRef.current) {

homeai-dashboard/src/lib/SchemaValidator.js

@@ -4,7 +4,43 @@ import schema from '../../schema/character.schema.json'
 const ajv = new Ajv({ allErrors: true, strict: false })
 const validate = ajv.compile(schema)
 
+/**
+ * Migrate a v1 character config to v2 in-place.
+ * Removes live2d/vtube fields, converts gaze_preset to gaze_presets array,
+ * and initialises new persona fields.
+ */
+export function migrateV1toV2(config) {
+  config.schema_version = 2
+
+  // Remove deprecated fields
+  delete config.live2d_expressions
+  delete config.vtube_ws_triggers
+
+  // Convert single gaze_preset string → gaze_presets array
+  if ('gaze_preset' in config) {
+    const old = config.gaze_preset
+    config.gaze_presets = old ? [{ preset: old, trigger: 'self-portrait' }] : []
+    delete config.gaze_preset
+  }
+  if (!config.gaze_presets) {
+    config.gaze_presets = []
+  }
+
+  // Initialise new fields if absent
+  if (config.background === undefined) config.background = ''
+  if (config.dialogue_style === undefined) config.dialogue_style = ''
+  if (config.appearance === undefined) config.appearance = ''
+  if (config.skills === undefined) config.skills = []
+
+  return config
+}
+
 export function validateCharacter(config) {
+  // Auto-migrate v1 → v2
+  if (config.schema_version === 1 || config.schema_version === undefined) {
+    migrateV1toV2(config)
+  }
+
   const valid = validate(config)
   if (!valid) {
     throw new Error(ajv.errorsText(validate.errors))

homeai-dashboard/src/lib/api.js

@@ -1,8 +1,30 @@
-export async function sendMessage(text) {
+const MAX_RETRIES = 3
-  const res = await fetch('/api/agent/message', {
+const RETRY_DELAY_MS = 2000
+
+async function fetchWithRetry(url, options, retries = MAX_RETRIES) {
+  for (let attempt = 1; attempt <= retries; attempt++) {
+    try {
+      const res = await fetch(url, options)
+      if (res.status === 502 && attempt < retries) {
+        // Bridge unreachable — wait and retry
+        await new Promise(r => setTimeout(r, RETRY_DELAY_MS * attempt))
+        continue
+      }
+      return res
+    } catch (err) {
+      if (attempt >= retries) throw err
+      await new Promise(r => setTimeout(r, RETRY_DELAY_MS * attempt))
+    }
+  }
+}
+
+export async function sendMessage(text, characterId = null) {
+  const payload = { message: text, agent: 'main' }
+  if (characterId) payload.character_id = characterId
+  const res = await fetchWithRetry('/api/agent/message', {
     method: 'POST',
     headers: { 'Content-Type': 'application/json' },
-    body: JSON.stringify({ message: text, agent: 'main' }),
+    body: JSON.stringify(payload),
   })
   if (!res.ok) {
     const err = await res.json().catch(() => ({ error: 'Request failed' }))
@@ -12,11 +34,13 @@ export async function sendMessage(text) {
   return data.response
 }
 
-export async function synthesize(text, voice) {
+export async function synthesize(text, voice, engine = 'kokoro', model = null) {
+  const payload = { text, voice, engine }
+  if (model) payload.model = model
   const res = await fetch('/api/tts', {
     method: 'POST',
     headers: { 'Content-Type': 'application/json' },
-    body: JSON.stringify({ text, voice }),
+    body: JSON.stringify(payload),
   })
   if (!res.ok) throw new Error('TTS failed')
   return await res.arrayBuffer()

homeai-dashboard/src/lib/constants.js

@@ -30,7 +30,15 @@ export const VOICES = [
   { id: 'bm_lewis', label: 'Lewis (M, UK)' },
 ]
 
+export const TTS_ENGINES = [
+  { id: 'kokoro', label: 'Kokoro (local)' },
+  { id: 'chatterbox', label: 'Chatterbox (voice clone)' },
+  { id: 'qwen3', label: 'Qwen3 TTS' },
+  { id: 'elevenlabs', label: 'ElevenLabs (cloud)' },
+]
+
 export const DEFAULT_SETTINGS = {
+  ttsEngine: 'kokoro',
   voice: DEFAULT_VOICE,
   autoTts: true,
   sttMode: 'bridge',

homeai-dashboard/src/lib/conversationApi.js

@@ -0,0 +1,25 @@
+export async function listConversations() {
+  const res = await fetch('/api/conversations')
+  if (!res.ok) throw new Error(`Failed to list conversations: ${res.status}`)
+  return res.json()
+}
+
+export async function getConversation(id) {
+  const res = await fetch(`/api/conversations/${encodeURIComponent(id)}`)
+  if (!res.ok) return null
+  return res.json()
+}
+
+export async function saveConversation(conversation) {
+  const res = await fetch('/api/conversations', {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(conversation),
+  })
+  if (!res.ok) throw new Error(`Failed to save conversation: ${res.status}`)
+}
+
+export async function deleteConversation(id) {
+  const res = await fetch(`/api/conversations/${encodeURIComponent(id)}`, { method: 'DELETE' })
+  if (!res.ok) throw new Error(`Failed to delete conversation: ${res.status}`)
+}

homeai-dashboard/src/lib/memoryApi.js

@@ -0,0 +1,45 @@
+export async function getPersonalMemories(characterId) {
+  const res = await fetch(`/api/memories/personal/${encodeURIComponent(characterId)}`)
+  if (!res.ok) return { characterId, memories: [] }
+  return res.json()
+}
+
+export async function savePersonalMemory(characterId, memory) {
+  const res = await fetch(`/api/memories/personal/${encodeURIComponent(characterId)}`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(memory),
+  })
+  if (!res.ok) throw new Error(`Failed to save memory: ${res.status}`)
+  return res.json()
+}
+
+export async function deletePersonalMemory(characterId, memoryId) {
+  const res = await fetch(`/api/memories/personal/${encodeURIComponent(characterId)}/${encodeURIComponent(memoryId)}`, {
+    method: 'DELETE',
+  })
+  if (!res.ok) throw new Error(`Failed to delete memory: ${res.status}`)
+}
+
+export async function getGeneralMemories() {
+  const res = await fetch('/api/memories/general')
+  if (!res.ok) return { memories: [] }
+  return res.json()
+}
+
+export async function saveGeneralMemory(memory) {
+  const res = await fetch('/api/memories/general', {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(memory),
+  })
+  if (!res.ok) throw new Error(`Failed to save memory: ${res.status}`)
+  return res.json()
+}
+
+export async function deleteGeneralMemory(memoryId) {
+  const res = await fetch(`/api/memories/general/${encodeURIComponent(memoryId)}`, {
+    method: 'DELETE',
+  })
+  if (!res.ok) throw new Error(`Failed to delete memory: ${res.status}`)
+}

homeai-dashboard/src/pages/Characters.jsx

@@ -1,23 +1,9 @@
-import { useState, useEffect } from 'react';
+import { useState, useEffect, useCallback } from 'react';
 import { useNavigate } from 'react-router-dom';
 import { validateCharacter } from '../lib/SchemaValidator';
 
-const STORAGE_KEY = 'homeai_characters';
 const ACTIVE_KEY = 'homeai_active_character';
 
-function loadProfiles() {
-  try {
-    const raw = localStorage.getItem(STORAGE_KEY);
-    return raw ? JSON.parse(raw) : [];
-  } catch {
-    return [];
-  }
-}
-
-function saveProfiles(profiles) {
-  localStorage.setItem(STORAGE_KEY, JSON.stringify(profiles));
-}
-
 function getActiveId() {
   return localStorage.getItem(ACTIVE_KEY) || null;
 }
@@ -27,15 +13,52 @@ function setActiveId(id) {
 }
 
 export default function Characters() {
-  const [profiles, setProfiles] = useState(loadProfiles);
+  const [profiles, setProfiles] = useState([]);
   const [activeId, setActive] = useState(getActiveId);
   const [error, setError] = useState(null);
   const [dragOver, setDragOver] = useState(false);
+  const [loading, setLoading] = useState(true);
+  const [satMap, setSatMap] = useState({ default: '', satellites: {} });
+  const [newSatId, setNewSatId] = useState('');
+  const [newSatChar, setNewSatChar] = useState('');
   const navigate = useNavigate();
 
+  // Load profiles and satellite map on mount
   useEffect(() => {
-    saveProfiles(profiles);
+    Promise.all([
-  }, [profiles]);
+      fetch('/api/characters').then(r => r.json()),
+      fetch('/api/satellite-map').then(r => r.json()),
+    ])
+      .then(([chars, map]) => {
+        setProfiles(chars);
+        setSatMap(map);
+        setLoading(false);
+      })
+      .catch(err => { setError(`Failed to load: ${err.message}`); setLoading(false); });
+  }, []);
+
+  const saveSatMap = useCallback(async (updated) => {
+    setSatMap(updated);
+    await fetch('/api/satellite-map', {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify(updated),
+    });
+  }, []);
+
+  const saveProfile = useCallback(async (profile) => {
+    const res = await fetch('/api/characters', {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify(profile),
+    });
+    if (!res.ok) throw new Error('Failed to save profile');
+  }, []);
+
+  const deleteProfile = useCallback(async (id) => {
+    const safeId = id.replace(/[^a-zA-Z0-9_\-\.]/g, '_');
+    await fetch(`/api/characters/${safeId}`, { method: 'DELETE' });
+  }, []);
 
   const handleImport = (e) => {
     const files = Array.from(e.target?.files || []);
@@ -47,12 +70,14 @@ export default function Characters() {
     files.forEach(file => {
       if (!file.name.endsWith('.json')) return;
       const reader = new FileReader();
-      reader.onload = (ev) => {
+      reader.onload = async (ev) => {
         try {
           const data = JSON.parse(ev.target.result);
           validateCharacter(data);
           const id = data.name + '_' + Date.now();
-          setProfiles(prev => [...prev, { id, data, image: null, addedAt: new Date().toISOString() }]);
+          const profile = { id, data, image: null, addedAt: new Date().toISOString() };
+          await saveProfile(profile);
+          setProfiles(prev => [...prev, profile]);
           setError(null);
         } catch (err) {
           setError(`Import failed for ${file.name}: ${err.message}`);
@@ -73,15 +98,17 @@ export default function Characters() {
     const file = e.target.files[0];
     if (!file) return;
     const reader = new FileReader();
-    reader.onload = (ev) => {
+    reader.onload = async (ev) => {
-      setProfiles(prev =>
+      const updated = profiles.map(p => p.id === profileId ? { ...p, image: ev.target.result } : p);
-        prev.map(p => p.id === profileId ? { ...p, image: ev.target.result } : p)
+      const profile = updated.find(p => p.id === profileId);
-      );
+      if (profile) await saveProfile(profile);
+      setProfiles(updated);
     };
     reader.readAsDataURL(file);
   };
 
-  const removeProfile = (id) => {
+  const removeProfile = async (id) => {
+    await deleteProfile(id);
     setProfiles(prev => prev.filter(p => p.id !== id));
     if (activeId === id) {
       setActive(null);
@@ -92,6 +119,28 @@ export default function Characters() {
   const activateProfile = (id) => {
     setActive(id);
     setActiveId(id);
+
+    // Sync active character's TTS settings to chat settings
+    const profile = profiles.find(p => p.id === id);
+    if (profile?.data?.tts) {
+      const tts = profile.data.tts;
+      const engine = tts.engine || 'kokoro';
+      let voice;
+      if (engine === 'kokoro') voice = tts.kokoro_voice || 'af_heart';
+      else if (engine === 'elevenlabs') voice = tts.elevenlabs_voice_id || '';
+      else if (engine === 'chatterbox') voice = tts.voice_ref_path || '';
+      else voice = '';
+
+      try {
+        const raw = localStorage.getItem('homeai_dashboard_settings');
+        const settings = raw ? JSON.parse(raw) : {};
+        localStorage.setItem('homeai_dashboard_settings', JSON.stringify({
+          ...settings,
+          ttsEngine: engine,
+          voice: voice,
+        }));
+      } catch { /* ignore */ }
+    }
   };
 
   const exportProfile = (profile) => {
@@ -125,13 +174,28 @@ export default function Characters() {
             )}
           </p>
         </div>
-        <label className="flex items-center gap-2 px-4 py-2 bg-indigo-600 hover:bg-indigo-500 text-white rounded-lg cursor-pointer transition-colors">
+        <div className="flex gap-3">
-          <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+          <button
-            <path strokeLinecap="round" strokeLinejoin="round" d="M12 4.5v15m7.5-7.5h-15" />
+            onClick={() => {
-          </svg>
+              sessionStorage.removeItem('edit_character');
-          Import JSON
+              sessionStorage.removeItem('edit_character_profile_id');
-          <input type="file" accept=".json" multiple className="hidden" onChange={handleImport} />
+              navigate('/editor');
-        </label>
+            }}
+            className="flex items-center gap-2 px-4 py-2 bg-indigo-600 hover:bg-indigo-500 text-white rounded-lg transition-colors"
+          >
+            <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+              <path strokeLinecap="round" strokeLinejoin="round" d="M12 4.5v15m7.5-7.5h-15" />
+            </svg>
+            New Character
+          </button>
+          <label className="flex items-center gap-2 px-4 py-2 bg-gray-800 hover:bg-gray-700 text-gray-300 rounded-lg cursor-pointer border border-gray-700 transition-colors">
+            <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+              <path strokeLinecap="round" strokeLinejoin="round" d="M3 16.5v2.25A2.25 2.25 0 005.25 21h13.5A2.25 2.25 0 0021 18.75V16.5m-13.5-9L12 3m0 0l4.5 4.5M12 3v13.5" />
+            </svg>
+            Import JSON
+            <input type="file" accept=".json" multiple className="hidden" onChange={handleImport} />
+          </label>
+        </div>
       </div>
 
       {error && (
@@ -158,7 +222,11 @@ export default function Characters() {
       </div>
 
       {/* Profile grid */}
-      {profiles.length === 0 ? (
+      {loading ? (
+        <div className="text-center py-16">
+          <p className="text-gray-500">Loading characters...</p>
+        </div>
+      ) : profiles.length === 0 ? (
         <div className="text-center py-16">
           <svg className="w-16 h-16 mx-auto text-gray-700 mb-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={1}>
             <path strokeLinecap="round" strokeLinejoin="round" d="M15.75 6a3.75 3.75 0 11-7.5 0 3.75 3.75 0 017.5 0zM4.501 20.118a7.5 7.5 0 0114.998 0A17.933 17.933 0 0112 21.75c-2.676 0-5.216-.584-7.499-1.632z" />
@@ -230,11 +298,32 @@ export default function Characters() {
                     <span className="px-2 py-0.5 bg-gray-700/70 text-gray-400 text-xs rounded-full">
                       {char.model_overrides?.primary || 'default'}
                     </span>
-                    {char.tts?.kokoro_voice && (
+                    {char.tts?.engine === 'kokoro' && char.tts?.kokoro_voice && (
                       <span className="px-2 py-0.5 bg-gray-700/70 text-gray-400 text-xs rounded-full">
                         {char.tts.kokoro_voice}
                       </span>
                     )}
+                    {char.tts?.engine === 'elevenlabs' && char.tts?.elevenlabs_voice_id && (
+                      <span className="px-2 py-0.5 bg-gray-700/70 text-gray-400 text-xs rounded-full" title={char.tts.elevenlabs_voice_id}>
+                        {char.tts.elevenlabs_voice_name || char.tts.elevenlabs_voice_id.slice(0, 8) + '…'}
+                      </span>
+                    )}
+                    {char.tts?.engine === 'chatterbox' && char.tts?.voice_ref_path && (
+                      <span className="px-2 py-0.5 bg-gray-700/70 text-gray-400 text-xs rounded-full" title={char.tts.voice_ref_path}>
+                        {char.tts.voice_ref_path.split('/').pop()}
+                      </span>
+                    )}
+                    {(() => {
+                      const defaultPreset = char.gaze_presets?.find(gp => gp.trigger === 'self-portrait')?.preset
+                        || char.gaze_presets?.[0]?.preset
+                        || char.gaze_preset
+                        || null;
+                      return defaultPreset ? (
+                        <span className="px-2 py-0.5 bg-violet-500/20 text-violet-300 text-xs rounded-full border border-violet-500/30" title={`GAZE: ${defaultPreset}`}>
+                          {defaultPreset}
+                        </span>
+                      ) : null;
+                    })()}
                   </div>
 
                   <div className="flex gap-2 pt-1">
@@ -287,6 +376,96 @@ export default function Characters() {
           })}
         </div>
       )}
+
+      {/* Satellite Assignment */}
+      {!loading && profiles.length > 0 && (
+        <div className="bg-gray-900 border border-gray-800 rounded-xl p-5 space-y-4">
+          <div>
+            <h2 className="text-lg font-semibold text-gray-200">Satellite Routing</h2>
+            <p className="text-xs text-gray-500 mt-1">Assign characters to voice satellites. Unmapped satellites use the default.</p>
+          </div>
+
+          {/* Default character */}
+          <div className="flex items-center gap-3">
+            <label className="text-sm text-gray-400 w-32 shrink-0">Default</label>
+            <select
+              value={satMap.default || ''}
+              onChange={(e) => saveSatMap({ ...satMap, default: e.target.value })}
+              className="flex-1 bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
+            >
+              <option value="">-- None --</option>
+              {profiles.map(p => (
+                <option key={p.id} value={p.id}>{p.data.display_name || p.data.name}</option>
+              ))}
+            </select>
+          </div>
+
+          {/* Per-satellite assignments */}
+          {Object.entries(satMap.satellites || {}).map(([satId, charId]) => (
+            <div key={satId} className="flex items-center gap-3">
+              <span className="text-sm text-gray-300 w-32 shrink-0 truncate font-mono" title={satId}>{satId}</span>
+              <select
+                value={charId}
+                onChange={(e) => {
+                  const updated = { ...satMap, satellites: { ...satMap.satellites, [satId]: e.target.value } };
+                  saveSatMap(updated);
+                }}
+                className="flex-1 bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
+              >
+                {profiles.map(p => (
+                  <option key={p.id} value={p.id}>{p.data.display_name || p.data.name}</option>
+                ))}
+              </select>
+              <button
+                onClick={() => {
+                  const { [satId]: _, ...rest } = satMap.satellites;
+                  saveSatMap({ ...satMap, satellites: rest });
+                }}
+                className="px-2 py-1.5 bg-gray-700 hover:bg-red-600 text-gray-400 hover:text-white rounded-lg transition-colors"
+                title="Remove"
+              >
+                <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+                  <path strokeLinecap="round" strokeLinejoin="round" d="M6 18L18 6M6 6l12 12" />
+                </svg>
+              </button>
+            </div>
+          ))}
+
+          {/* Add new satellite */}
+          <div className="flex items-center gap-3 pt-2 border-t border-gray-800">
+            <input
+              type="text"
+              value={newSatId}
+              onChange={(e) => setNewSatId(e.target.value)}
+              placeholder="Satellite ID (from bridge log)"
+              className="w-32 shrink-0 bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500 font-mono"
+            />
+            <select
+              value={newSatChar}
+              onChange={(e) => setNewSatChar(e.target.value)}
+              className="flex-1 bg-gray-800 text-gray-200 text-sm rounded-lg px-3 py-2 border border-gray-700 focus:outline-none focus:border-indigo-500"
+            >
+              <option value="">-- Select Character --</option>
+              {profiles.map(p => (
+                <option key={p.id} value={p.id}>{p.data.display_name || p.data.name}</option>
+              ))}
+            </select>
+            <button
+              onClick={() => {
+                if (newSatId && newSatChar) {
+                  saveSatMap({ ...satMap, satellites: { ...satMap.satellites, [newSatId]: newSatChar } });
+                  setNewSatId('');
+                  setNewSatChar('');
+                }
+              }}
+              disabled={!newSatId || !newSatChar}
+              className="px-3 py-1.5 bg-indigo-600 hover:bg-indigo-500 disabled:bg-gray-700 disabled:text-gray-500 text-white text-sm rounded-lg transition-colors"
+            >
+              Add
+            </button>
+          </div>
+        </div>
+      )}
     </div>
   );
 }

homeai-dashboard/src/pages/Chat.jsx

@@ -1,115 +1,146 @@
-import { useState, useEffect, useCallback } from 'react'
+import { useState, useCallback } from 'react'
 import ChatPanel from '../components/ChatPanel'
 import InputBar from '../components/InputBar'
 import StatusIndicator from '../components/StatusIndicator'
 import SettingsDrawer from '../components/SettingsDrawer'
+import ConversationList from '../components/ConversationList'
 import { useSettings } from '../hooks/useSettings'
 import { useBridgeHealth } from '../hooks/useBridgeHealth'
 import { useChat } from '../hooks/useChat'
 import { useTtsPlayback } from '../hooks/useTtsPlayback'
 import { useVoiceInput } from '../hooks/useVoiceInput'
+import { useActiveCharacter } from '../hooks/useActiveCharacter'
+import { useConversations } from '../hooks/useConversations'
 
 export default function Chat() {
   const { settings, updateSetting } = useSettings()
   const isOnline = useBridgeHealth()
-  const { messages, isLoading, send, clearHistory } = useChat()
+  const character = useActiveCharacter()
-  const { isPlaying, speak, stop } = useTtsPlayback(settings.voice)
+  const {
+    conversations, activeId, isLoading: isLoadingList,
+    select, create, remove, updateMeta,
+  } = useConversations()
+
+  const convMeta = {
+    characterId: character?.id || '',
+    characterName: character?.name || '',
+  }
+
+  const { messages, isLoading, isLoadingConv, send, clearHistory } = useChat(activeId, convMeta, updateMeta)
+
+  // Use character's TTS config if available, fall back to global settings
+  const ttsEngine = character?.tts?.engine || settings.ttsEngine
+  const ttsVoice = ttsEngine === 'elevenlabs'
+    ? (character?.tts?.elevenlabs_voice_id || settings.voice)
+    : (character?.tts?.kokoro_voice || settings.voice)
+  const ttsModel = ttsEngine === 'elevenlabs' ? (character?.tts?.elevenlabs_model || null) : null
+  const { isPlaying, speak, stop } = useTtsPlayback(ttsVoice, ttsEngine, ttsModel)
   const { isRecording, isTranscribing, startRecording, stopRecording } = useVoiceInput(settings.sttMode)
   const [settingsOpen, setSettingsOpen] = useState(false)
 
-  // Send a message and optionally speak the response
   const handleSend = useCallback(async (text) => {
-    const response = await send(text)
+    // Auto-create a conversation if none is active
+    let newId = null
+    if (!activeId) {
+      newId = await create(convMeta.characterId, convMeta.characterName)
+    }
+    const response = await send(text, newId)
     if (response && settings.autoTts) {
       speak(response)
     }
-  }, [send, settings.autoTts, speak])
+  }, [activeId, create, convMeta, send, settings.autoTts, speak])
 
-  // Toggle voice recording
   const handleVoiceToggle = useCallback(async () => {
     if (isRecording) {
       const text = await stopRecording()
-      if (text) {
+      if (text) handleSend(text)
-        handleSend(text)
-      }
     } else {
       startRecording()
     }
   }, [isRecording, stopRecording, startRecording, handleSend])
 
-  // Space bar push-to-talk when input not focused
+  const handleNewChat = useCallback(() => {
-  useEffect(() => {
+    create(convMeta.characterId, convMeta.characterName)
-    const handleKeyDown = (e) => {
+  }, [create, convMeta])
-      if (e.code === 'Space' && e.target.tagName !== 'TEXTAREA' && e.target.tagName !== 'INPUT') {
-        e.preventDefault()
-        handleVoiceToggle()
-      }
-    }
-    window.addEventListener('keydown', handleKeyDown)
-    return () => window.removeEventListener('keydown', handleKeyDown)
-  }, [handleVoiceToggle])
 
   return (
-    <div className="flex-1 flex flex-col min-h-0">
+    <div className="flex-1 flex min-h-0">
-      {/* Status bar */}
+      {/* Conversation sidebar */}
-      <header className="flex items-center justify-between px-4 py-2 border-b border-gray-800/50 shrink-0">
+      <ConversationList
-        <div className="flex items-center gap-2">
+        conversations={conversations}
-          <StatusIndicator isOnline={isOnline} />
+        activeId={activeId}
-          <span className="text-xs text-gray-500">
+        onCreate={handleNewChat}
-            {isOnline === null ? 'Connecting...' : isOnline ? 'Connected' : 'Offline'}
+        onSelect={select}
-          </span>
+        onDelete={remove}
-        </div>
+      />
-        <div className="flex items-center gap-2">
-          {messages.length > 0 && (
-            <button
-              onClick={clearHistory}
-              className="text-xs text-gray-500 hover:text-gray-300 transition-colors px-2 py-1"
-              title="Clear conversation"
-            >
-              Clear
-            </button>
-          )}
-          {isPlaying && (
-            <button
-              onClick={stop}
-              className="text-xs text-indigo-400 hover:text-indigo-300 transition-colors px-2 py-1"
-              title="Stop speaking"
-            >
-              Stop audio
-            </button>
-          )}
-          <button
-            onClick={() => setSettingsOpen(true)}
-            className="text-gray-500 hover:text-gray-300 transition-colors p-1"
-            title="Settings"
-          >
-            <svg className="w-5 h-5" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={1.5}>
-              <path strokeLinecap="round" strokeLinejoin="round" d="M9.594 3.94c.09-.542.56-.94 1.11-.94h2.593c.55 0 1.02.398 1.11.94l.213 1.281c.063.374.313.686.645.87.074.04.147.083.22.127.325.196.72.257 1.075.124l1.217-.456a1.125 1.125 0 011.37.49l1.296 2.247a1.125 1.125 0 01-.26 1.431l-1.003.827c-.293.241-.438.613-.43.992a7.723 7.723 0 010 .255c-.008.378.137.75.43.991l1.004.827c.424.35.534.955.26 1.43l-1.298 2.247a1.125 1.125 0 01-1.369.491l-1.217-.456c-.355-.133-.75-.072-1.076.124a6.47 6.47 0 01-.22.128c-.331.183-.581.495-.644.869l-.213 1.281c-.09.543-.56.941-1.11.941h-2.594c-.55 0-1.019-.398-1.11-.94l-.213-1.281c-.062-.374-.312-.686-.644-.87a6.52 6.52 0 01-.22-.127c-.325-.196-.72-.257-1.076-.124l-1.217.456a1.125 1.125 0 01-1.369-.49l-1.297-2.247a1.125 1.125 0 01.26-1.431l1.004-.827c.292-.24.437-.613.43-.991a6.932 6.932 0 010-.255c.007-.38-.138-.751-.43-.992l-1.004-.827a1.125 1.125 0 01-.26-1.43l1.297-2.247a1.125 1.125 0 011.37-.491l1.216.456c.356.133.751.072 1.076-.124.072-.044.146-.086.22-.128.332-.183.582-.495.644-.869l.214-1.28z" />
-              <path strokeLinecap="round" strokeLinejoin="round" d="M15 12a3 3 0 11-6 0 3 3 0 016 0z" />
-            </svg>
-          </button>
-        </div>
-      </header>
 
       {/* Chat area */}
-      <ChatPanel messages={messages} isLoading={isLoading} onReplay={speak} />
+      <div className="flex-1 flex flex-col min-h-0 min-w-0">
+        {/* Status bar */}
+        <header className="flex items-center justify-between px-4 py-2 border-b border-gray-800/50 shrink-0">
+          <div className="flex items-center gap-2">
+            <StatusIndicator isOnline={isOnline} />
+            <span className="text-xs text-gray-500">
+              {isOnline === null ? 'Connecting...' : isOnline ? 'Connected' : 'Offline'}
+            </span>
+          </div>
+          <div className="flex items-center gap-2">
+            {messages.length > 0 && (
+              <button
+                onClick={clearHistory}
+                className="text-xs text-gray-500 hover:text-gray-300 transition-colors px-2 py-1"
+                title="Clear conversation"
+              >
+                Clear
+              </button>
+            )}
+            {isPlaying && (
+              <button
+                onClick={stop}
+                className="text-xs text-indigo-400 hover:text-indigo-300 transition-colors px-2 py-1"
+                title="Stop speaking"
+              >
+                Stop audio
+              </button>
+            )}
+            <button
+              onClick={() => setSettingsOpen(true)}
+              className="text-gray-500 hover:text-gray-300 transition-colors p-1"
+              title="Settings"
+            >
+              <svg className="w-5 h-5" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={1.5}>
+                <path strokeLinecap="round" strokeLinejoin="round" d="M9.594 3.94c.09-.542.56-.94 1.11-.94h2.593c.55 0 1.02.398 1.11.94l.213 1.281c.063.374.313.686.645.87.074.04.147.083.22.127.325.196.72.257 1.075.124l1.217-.456a1.125 1.125 0 011.37.49l1.296 2.247a1.125 1.125 0 01-.26 1.431l-1.003.827c-.293.241-.438.613-.43.992a7.723 7.723 0 010 .255c-.008.378.137.75.43.991l1.004.827c.424.35.534.955.26 1.43l-1.298 2.247a1.125 1.125 0 01-1.369.491l-1.217-.456c-.355-.133-.75-.072-1.076.124a6.47 6.47 0 01-.22.128c-.331.183-.581.495-.644.869l-.213 1.281c-.09.543-.56.941-1.11.941h-2.594c-.55 0-1.019-.398-1.11-.94l-.213-1.281c-.062-.374-.312-.686-.644-.87a6.52 6.52 0 01-.22-.127c-.325-.196-.72-.257-1.076-.124l-1.217.456a1.125 1.125 0 01-1.369-.49l-1.297-2.247a1.125 1.125 0 01.26-1.431l1.004-.827c.292-.24.437-.613.43-.991a6.932 6.932 0 010-.255c.007-.38-.138-.751-.43-.992l-1.004-.827a1.125 1.125 0 01-.26-1.43l1.297-2.247a1.125 1.125 0 011.37-.491l1.216.456c.356.133.751.072 1.076-.124.072-.044.146-.086.22-.128.332-.183.582-.495.644-.869l.214-1.28z" />
+                <path strokeLinecap="round" strokeLinejoin="round" d="M15 12a3 3 0 11-6 0 3 3 0 016 0z" />
+              </svg>
+            </button>
+          </div>
+        </header>
 
-      {/* Input */}
+        {/* Messages */}
-      <InputBar
+        <ChatPanel
-        onSend={handleSend}
+          messages={messages}
-        onVoiceToggle={handleVoiceToggle}
+          isLoading={isLoading || isLoadingConv}
-        isLoading={isLoading}
+          onReplay={speak}
-        isRecording={isRecording}
+          character={character}
-        isTranscribing={isTranscribing}
+        />
-      />
 
-      {/* Settings drawer */}
+        {/* Input */}
-      <SettingsDrawer
+        <InputBar
-        isOpen={settingsOpen}
+          onSend={handleSend}
-        onClose={() => setSettingsOpen(false)}
+          onVoiceToggle={handleVoiceToggle}
-        settings={settings}
+          isLoading={isLoading}
-        onUpdate={updateSetting}
+          isRecording={isRecording}
-      />
+          isTranscribing={isTranscribing}
+        />
+
+        {/* Settings drawer */}
+        <SettingsDrawer
+          isOpen={settingsOpen}
+          onClose={() => setSettingsOpen(false)}
+          settings={settings}
+          onUpdate={updateSetting}
+        />
+      </div>
     </div>
   )
 }

homeai-dashboard/src/pages/Editor.jsx

@@ -1,14 +1,18 @@
 import React, { useState, useEffect, useRef } from 'react';
-import { validateCharacter } from '../lib/SchemaValidator';
+import { validateCharacter, migrateV1toV2 } from '../lib/SchemaValidator';
 
 const DEFAULT_CHARACTER = {
-  schema_version: 1,
+  schema_version: 2,
-  name: "aria",
+  name: "",
-  display_name: "Aria",
+  display_name: "",
-  description: "Default HomeAI assistant persona",
+  description: "",
-  system_prompt: "You are Aria, a warm, curious, and helpful AI assistant living in the home. You speak naturally and conversationally — never robotic. You are knowledgeable but never condescending. You remember the people you live with and build on those memories over time. Keep responses concise when controlling smart home devices; be more expressive in casual conversation. Never break character.",
+  background: "",
+  dialogue_style: "",
+  appearance: "",
+  skills: [],
+  system_prompt: "",
   model_overrides: {
-    primary: "llama3.3:70b",
+    primary: "qwen3.5:35b-a3b",
     fast: "qwen2.5:7b"
   },
   tts: {
@@ -16,24 +20,8 @@ const DEFAULT_CHARACTER = {
     kokoro_voice: "af_heart",
     speed: 1.0
   },
-  live2d_expressions: {
+  gaze_presets: [],
-    idle: "expr_idle",
+  custom_rules: [],
-    listening: "expr_listening",
-    thinking: "expr_thinking",
-    speaking: "expr_speaking",
-    happy: "expr_happy",
-    sad: "expr_sad",
-    surprised: "expr_surprised",
-    error: "expr_error"
-  },
-  vtube_ws_triggers: {
-    thinking: { type: "hotkey", id: "expr_thinking" },
-    speaking: { type: "hotkey", id: "expr_speaking" },
-    idle: { type: "hotkey", id: "expr_idle" }
-  },
-  custom_rules: [
-    { trigger: "good morning", response: "Good morning! How did you sleep?", condition: "time_of_day == morning" }
-  ],
   notes: ""
 };
 
@@ -43,7 +31,12 @@ export default function Editor() {
     if (editData) {
       sessionStorage.removeItem('edit_character');
       try {
-        return JSON.parse(editData);
+        const parsed = JSON.parse(editData);
+        // Auto-migrate v1 data
+        if (parsed.schema_version === 1 || !parsed.schema_version) {
+          migrateV1toV2(parsed);
+        }
+        return parsed;
       } catch {
         return DEFAULT_CHARACTER;
       }
@@ -52,6 +45,7 @@ export default function Editor() {
   });
   const [error, setError] = useState(null);
   const [saved, setSaved] = useState(false);
+  const isEditing = !!sessionStorage.getItem('edit_character_profile_id');
 
   // TTS preview state
   const [ttsState, setTtsState] = useState('idle');
@@ -65,6 +59,19 @@ export default function Editor() {
   const [elevenLabsModels, setElevenLabsModels] = useState([]);
   const [isLoadingElevenLabs, setIsLoadingElevenLabs] = useState(false);
 
+  // GAZE presets state (from API)
+  const [availableGazePresets, setAvailableGazePresets] = useState([]);
+  const [isLoadingGaze, setIsLoadingGaze] = useState(false);
+
+  // Character lookup state
+  const [lookupName, setLookupName] = useState('');
+  const [lookupFranchise, setLookupFranchise] = useState('');
+  const [isLookingUp, setIsLookingUp] = useState(false);
+  const [lookupDone, setLookupDone] = useState(false);
+
+  // Skills input state
+  const [newSkill, setNewSkill] = useState('');
+
   const fetchElevenLabsData = async (key) => {
     if (!key) return;
     setIsLoadingElevenLabs(true);
@@ -95,6 +102,16 @@ export default function Editor() {
     }
   }, [character.tts.engine]);
 
+  // Fetch GAZE presets on mount
+  useEffect(() => {
+    setIsLoadingGaze(true);
+    fetch('/api/gaze/presets')
+      .then(r => r.ok ? r.json() : { presets: [] })
+      .then(data => setAvailableGazePresets(data.presets || []))
+      .catch(() => {})
+      .finally(() => setIsLoadingGaze(false));
+  }, []);
+
   useEffect(() => {
     return () => {
       if (audioRef.current) { audioRef.current.pause(); audioRef.current = null; }
@@ -119,27 +136,35 @@ export default function Editor() {
     }
   };
 
-  const handleSaveToProfiles = () => {
+  const handleSaveToProfiles = async () => {
     try {
       validateCharacter(character);
       setError(null);
 
       const profileId = sessionStorage.getItem('edit_character_profile_id');
-      const storageKey = 'homeai_characters';
+      let profile;
-      const raw = localStorage.getItem(storageKey);
-      let profiles = raw ? JSON.parse(raw) : [];
 
       if (profileId) {
-        profiles = profiles.map(p =>
+        const res = await fetch('/api/characters');
-          p.id === profileId ? { ...p, data: character } : p
+        const profiles = await res.json();
-        );
+        const existing = profiles.find(p => p.id === profileId);
-        sessionStorage.removeItem('edit_character_profile_id');
+        profile = existing
+          ? { ...existing, data: character }
+          : { id: profileId, data: character, image: null, addedAt: new Date().toISOString() };
+        // Keep the profile ID in sessionStorage so subsequent saves update the same file
       } else {
         const id = character.name + '_' + Date.now();
-        profiles.push({ id, data: character, image: null, addedAt: new Date().toISOString() });
+        profile = { id, data: character, image: null, addedAt: new Date().toISOString() };
+        // Store the new ID so subsequent saves update the same file
+        sessionStorage.setItem('edit_character_profile_id', profile.id);
       }
 
-      localStorage.setItem(storageKey, JSON.stringify(profiles));
+      await fetch('/api/characters', {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify(profile),
+      });
+
       setSaved(true);
       setTimeout(() => setSaved(false), 2000);
     } catch (err) {
@@ -164,6 +189,59 @@ export default function Editor() {
     reader.readAsText(file);
   };
 
+  // Character lookup from MCP
+  const handleCharacterLookup = async () => {
+    if (!lookupName || !lookupFranchise) return;
+    setIsLookingUp(true);
+    setError(null);
+    try {
+      const res = await fetch('/api/character-lookup', {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify({ name: lookupName, franchise: lookupFranchise }),
+      });
+      if (!res.ok) {
+        const err = await res.json().catch(() => ({ error: 'Lookup failed' }));
+        throw new Error(err.error || `Lookup returned ${res.status}`);
+      }
+      const data = await res.json();
+
+      // Build dialogue_style from personality + notable quotes
+      let dialogueStyle = data.personality || '';
+      if (data.notable_quotes?.length) {
+        dialogueStyle += '\n\nExample dialogue:\n' + data.notable_quotes.map(q => `"${q}"`).join('\n');
+      }
+
+      // Filter abilities to clean text-only entries (skip image captions)
+      const skills = (data.abilities || [])
+        .filter(a => a.length > 20 && !a.includes('.jpg') && !a.includes('.png'))
+        .slice(0, 10);
+
+      // Auto-generate system prompt
+      const promptName = character.display_name || lookupName;
+      const personality = data.personality ? data.personality.split('.').slice(0, 3).join('.') + '.' : '';
+      const systemPrompt = `You are ${promptName} from ${lookupFranchise}. ${personality} Stay in character at all times. Respond naturally and conversationally.`;
+
+      setCharacter(prev => ({
+        ...prev,
+        name: prev.name || lookupName.toLowerCase().replace(/\s+/g, '_'),
+        display_name: prev.display_name || lookupName,
+        description: data.description ? data.description.split('.').slice(0, 2).join('.') + '.' : prev.description,
+        background: data.background || prev.background,
+        appearance: data.appearance || prev.appearance,
+        dialogue_style: dialogueStyle || prev.dialogue_style,
+        skills: skills.length > 0 ? skills : prev.skills,
+        system_prompt: prev.system_prompt || systemPrompt,
+      }));
+
+      setLookupDone(true);
+    } catch (err) {
+      setError(`Character lookup failed: ${err.message}`);
+    } finally {
+      setIsLookingUp(false);
+    }
+  };
+
   const handleChange = (field, value) => {
     setCharacter(prev => ({ ...prev, [field]: value }));
   };
@@ -175,6 +253,50 @@ export default function Editor() {
     }));
   };
 
+  // Skills helpers
+  const addSkill = () => {
+    const trimmed = newSkill.trim();
+    if (!trimmed) return;
+    setCharacter(prev => ({
+      ...prev,
+      skills: [...(prev.skills || []), trimmed]
+    }));
+    setNewSkill('');
+  };
+
+  const removeSkill = (index) => {
+    setCharacter(prev => {
+      const updated = [...(prev.skills || [])];
+      updated.splice(index, 1);
+      return { ...prev, skills: updated };
+    });
+  };
+
+  // GAZE preset helpers
+  const addGazePreset = () => {
+    setCharacter(prev => ({
+      ...prev,
+      gaze_presets: [...(prev.gaze_presets || []), { preset: '', trigger: 'self-portrait' }]
+    }));
+  };
+
+  const removeGazePreset = (index) => {
+    setCharacter(prev => {
+      const updated = [...(prev.gaze_presets || [])];
+      updated.splice(index, 1);
+      return { ...prev, gaze_presets: updated };
+    });
+  };
+
+  const handleGazePresetChange = (index, field, value) => {
+    setCharacter(prev => {
+      const updated = [...(prev.gaze_presets || [])];
+      updated[index] = { ...updated[index], [field]: value };
+      return { ...prev, gaze_presets: updated };
+    });
+  };
+
+  // Custom rules helpers
   const handleRuleChange = (index, field, value) => {
     setCharacter(prev => {
       const newRules = [...(prev.custom_rules || [])];
@@ -198,37 +320,40 @@ export default function Editor() {
     });
   };
 
+  // TTS preview
   const stopPreview = () => {
-    if (audioRef.current) {
+    if (audioRef.current) { audioRef.current.pause(); audioRef.current = null; }
-      audioRef.current.pause();
+    if (objectUrlRef.current) { URL.revokeObjectURL(objectUrlRef.current); objectUrlRef.current = null; }
-      audioRef.current = null;
-    }
-    if (objectUrlRef.current) {
-      URL.revokeObjectURL(objectUrlRef.current);
-      objectUrlRef.current = null;
-    }
     window.speechSynthesis.cancel();
     setTtsState('idle');
   };
 
   const previewTTS = async () => {
     stopPreview();
-    const text = previewText || `Hi, I am ${character.display_name}. This is a preview of my voice.`;
+    const text = previewText || `Hi, I am ${character.display_name || character.name}. This is a preview of my voice.`;
+    const engine = character.tts.engine;
 
-    if (character.tts.engine === 'kokoro') {
+    let bridgeBody = null;
+    if (engine === 'kokoro') {
+      bridgeBody = { text, voice: character.tts.kokoro_voice, engine: 'kokoro' };
+    } else if (engine === 'elevenlabs' && character.tts.elevenlabs_voice_id) {
+      bridgeBody = { text, voice: character.tts.elevenlabs_voice_id, engine: 'elevenlabs', model: character.tts.elevenlabs_model };
+    }
+
+    if (bridgeBody) {
       setTtsState('loading');
       let blob;
       try {
         const response = await fetch('/api/tts', {
           method: 'POST',
           headers: { 'Content-Type': 'application/json' },
-          body: JSON.stringify({ text, voice: character.tts.kokoro_voice })
+          body: JSON.stringify(bridgeBody)
         });
         if (!response.ok) throw new Error('TTS bridge returned ' + response.status);
         blob = await response.blob();
       } catch (err) {
         setTtsState('idle');
-        setError(`Kokoro preview failed: ${err.message}. Falling back to browser TTS.`);
+        setError(`${engine} preview failed: ${err.message}. Falling back to browser TTS.`);
         runBrowserTTS(text);
         return;
       }
@@ -269,7 +394,9 @@ export default function Editor() {
         <div>
           <h1 className="text-3xl font-bold text-gray-100">Character Editor</h1>
           <p className="text-sm text-gray-500 mt-1">
-            Editing: {character.display_name || character.name}
+            {character.display_name || character.name
+              ? `Editing: ${character.display_name || character.name}`
+              : 'New character'}
           </p>
         </div>
         <div className="flex gap-3">
@@ -311,6 +438,64 @@ export default function Editor() {
       {error && (
         <div className="bg-red-900/30 border border-red-500/50 text-red-300 px-4 py-3 rounded-lg text-sm">
           {error}
+          <button onClick={() => setError(null)} className="ml-2 text-red-400 hover:text-red-300">&times;</button>
+        </div>
+      )}
+
+      {/* Character Lookup — auto-fill from fictional character wiki */}
+      {!isEditing && (
+        <div className={cardClass}>
+          <div className="flex items-center gap-2">
+            <svg className="w-5 h-5 text-indigo-400" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+              <path strokeLinecap="round" strokeLinejoin="round" d="M21 21l-5.197-5.197m0 0A7.5 7.5 0 105.196 5.196a7.5 7.5 0 0010.607 10.607z" />
+            </svg>
+            <h2 className="text-lg font-semibold text-gray-200">Auto-fill from Character</h2>
+          </div>
+          <p className="text-xs text-gray-500">Fetch character data from Fandom/Wikipedia to auto-populate fields. You can edit everything after.</p>
+          <div className="flex gap-3 items-end">
+            <div className="flex-1">
+              <label className={labelClass}>Character Name</label>
+              <input
+                type="text"
+                className={inputClass}
+                value={lookupName}
+                onChange={(e) => setLookupName(e.target.value)}
+                placeholder="e.g. Tifa Lockhart"
+              />
+            </div>
+            <div className="flex-1">
+              <label className={labelClass}>Franchise / Series</label>
+              <input
+                type="text"
+                className={inputClass}
+                value={lookupFranchise}
+                onChange={(e) => setLookupFranchise(e.target.value)}
+                placeholder="e.g. Final Fantasy VII"
+              />
+            </div>
+            <button
+              onClick={handleCharacterLookup}
+              disabled={isLookingUp || !lookupName || !lookupFranchise}
+              className={`flex items-center gap-2 px-5 py-2 rounded-lg text-white transition-colors whitespace-nowrap ${
+                isLookingUp
+                  ? 'bg-indigo-800 cursor-wait'
+                  : lookupDone
+                    ? 'bg-emerald-600 hover:bg-emerald-500'
+                    : 'bg-indigo-600 hover:bg-indigo-500 disabled:bg-gray-700 disabled:text-gray-500'
+              }`}
+            >
+              {isLookingUp && (
+                <svg className="w-4 h-4 animate-spin" viewBox="0 0 24 24" fill="none">
+                  <circle className="opacity-25" cx="12" cy="12" r="10" stroke="currentColor" strokeWidth="4" />
+                  <path className="opacity-75" fill="currentColor" d="M4 12a8 8 0 018-8V0C5.373 0 0 5.373 0 12h4z" />
+                </svg>
+              )}
+              {isLookingUp ? 'Fetching...' : lookupDone ? 'Fetched' : 'Lookup'}
+            </button>
+          </div>
+          {lookupDone && (
+            <p className="text-xs text-emerald-400">Fields populated from wiki data. Review and edit below.</p>
+          )}
         </div>
       )}
 
@@ -324,11 +509,11 @@ export default function Editor() {
           </div>
           <div>
             <label className={labelClass}>Display Name</label>
-            <input type="text" className={inputClass} value={character.display_name} onChange={(e) => handleChange('display_name', e.target.value)} />
+            <input type="text" className={inputClass} value={character.display_name || ''} onChange={(e) => handleChange('display_name', e.target.value)} />
           </div>
           <div>
             <label className={labelClass}>Description</label>
-            <input type="text" className={inputClass} value={character.description} onChange={(e) => handleChange('description', e.target.value)} />
+            <input type="text" className={inputClass} value={character.description || ''} onChange={(e) => handleChange('description', e.target.value)} />
           </div>
         </div>
 
@@ -359,7 +544,14 @@ export default function Editor() {
               <div>
                 <label className={labelClass}>Voice ID</label>
                 {elevenLabsVoices.length > 0 ? (
-                  <select className={selectClass} value={character.tts.elevenlabs_voice_id || ''} onChange={(e) => handleNestedChange('tts', 'elevenlabs_voice_id', e.target.value)}>
+                  <select className={selectClass} value={character.tts.elevenlabs_voice_id || ''} onChange={(e) => {
+                    const voiceId = e.target.value;
+                    const voice = elevenLabsVoices.find(v => v.voice_id === voiceId);
+                    setCharacter(prev => ({
+                      ...prev,
+                      tts: { ...prev.tts, elevenlabs_voice_id: voiceId, elevenlabs_voice_name: voice?.name || '' }
+                    }));
+                  }}>
                     <option value="">-- Select Voice --</option>
                     {elevenLabsVoices.map(v => (
                       <option key={v.voice_id} value={v.voice_id}>{v.name} ({v.category})</option>
@@ -439,7 +631,7 @@ export default function Editor() {
               className={inputClass}
               value={previewText}
               onChange={(e) => setPreviewText(e.target.value)}
-              placeholder={`Hi, I am ${character.display_name}. This is a preview of my voice.`}
+              placeholder={`Hi, I am ${character.display_name || character.name || 'your character'}. This is a preview of my voice.`}
             />
           </div>
           <div className="flex gap-2">
@@ -474,7 +666,9 @@ export default function Editor() {
           <p className="text-xs text-gray-600">
             {character.tts.engine === 'kokoro'
               ? 'Previews via local Kokoro TTS bridge (port 8081).'
-              : 'Uses browser TTS for preview. Local TTS available with Kokoro engine.'}
+              : character.tts.engine === 'elevenlabs'
+                ? 'Previews via ElevenLabs through bridge.'
+                : 'Uses browser TTS for preview. Local TTS available with Kokoro engine.'}
           </p>
         </div>
       </div>
@@ -483,25 +677,154 @@ export default function Editor() {
       <div className={cardClass}>
         <div className="flex justify-between items-center">
           <h2 className="text-lg font-semibold text-gray-200">System Prompt</h2>
-          <span className="text-xs text-gray-600">{character.system_prompt.length} chars</span>
+          <span className="text-xs text-gray-600">{(character.system_prompt || '').length} chars</span>
         </div>
         <textarea
           className={inputClass + " h-32 resize-y"}
           value={character.system_prompt}
           onChange={(e) => handleChange('system_prompt', e.target.value)}
+          placeholder="You are [character name]. Describe their personality, behaviour, and role..."
         />
       </div>
 
+      {/* Character Profile — new v2 fields */}
+      <div className={cardClass}>
+        <h2 className="text-lg font-semibold text-gray-200">Character Profile</h2>
+        <div>
+          <label className={labelClass}>Background / Backstory</label>
+          <textarea
+            className={inputClass + " h-28 resize-y text-sm"}
+            value={character.background || ''}
+            onChange={(e) => handleChange('background', e.target.value)}
+            placeholder="Character history, origins, key life events..."
+          />
+        </div>
+        <div>
+          <label className={labelClass}>Appearance</label>
+          <textarea
+            className={inputClass + " h-24 resize-y text-sm"}
+            value={character.appearance || ''}
+            onChange={(e) => handleChange('appearance', e.target.value)}
+            placeholder="Physical description — also used for image generation prompts..."
+          />
+        </div>
+        <div>
+          <label className={labelClass}>Dialogue Style & Examples</label>
+          <textarea
+            className={inputClass + " h-24 resize-y text-sm"}
+            value={character.dialogue_style || ''}
+            onChange={(e) => handleChange('dialogue_style', e.target.value)}
+            placeholder="How the persona speaks, their tone, mannerisms, and example lines..."
+          />
+        </div>
+        <div>
+          <label className={labelClass}>Skills & Interests</label>
+          <div className="flex flex-wrap gap-2 mb-2">
+            {(character.skills || []).map((skill, idx) => (
+              <span
+                key={idx}
+                className="inline-flex items-center gap-1 px-3 py-1 bg-indigo-500/20 text-indigo-300 text-sm rounded-full border border-indigo-500/30"
+              >
+                {skill.length > 80 ? skill.slice(0, 80) + '...' : skill}
+                <button
+                  onClick={() => removeSkill(idx)}
+                  className="ml-1 text-indigo-400 hover:text-red-400 transition-colors"
+                >
+                  <svg className="w-3 h-3" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={3}>
+                    <path strokeLinecap="round" strokeLinejoin="round" d="M6 18L18 6M6 6l12 12" />
+                  </svg>
+                </button>
+              </span>
+            ))}
+          </div>
+          <div className="flex gap-2">
+            <input
+              type="text"
+              className={inputClass + " text-sm"}
+              value={newSkill}
+              onChange={(e) => setNewSkill(e.target.value)}
+              onKeyDown={(e) => { if (e.key === 'Enter') { e.preventDefault(); addSkill(); } }}
+              placeholder="Add a skill or interest..."
+            />
+            <button
+              onClick={addSkill}
+              disabled={!newSkill.trim()}
+              className="px-3 py-2 bg-indigo-600 hover:bg-indigo-500 disabled:bg-gray-700 disabled:text-gray-500 text-white text-sm rounded-lg transition-colors whitespace-nowrap"
+            >
+              Add
+            </button>
+          </div>
+        </div>
+      </div>
+
       <div className="grid grid-cols-1 md:grid-cols-2 gap-6">
-        {/* Live2D Expressions */}
+        {/* Image Generation — GAZE presets */}
         <div className={cardClass}>
-          <h2 className="text-lg font-semibold text-gray-200">Live2D Expressions</h2>
+          <div className="flex justify-between items-center">
-          {Object.entries(character.live2d_expressions).map(([key, val]) => (
+            <h2 className="text-lg font-semibold text-gray-200">GAZE Presets</h2>
-            <div key={key} className="flex justify-between items-center gap-4">
+            <button onClick={addGazePreset} className="flex items-center gap-1 bg-indigo-600 hover:bg-indigo-500 text-white px-3 py-1.5 rounded-lg text-sm transition-colors">
-              <label className="text-sm font-medium text-gray-400 w-1/3 capitalize">{key}</label>
+              <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
-              <input type="text" className={inputClass + " w-2/3"} value={val} onChange={(e) => handleNestedChange('live2d_expressions', key, e.target.value)} />
+                <path strokeLinecap="round" strokeLinejoin="round" d="M12 4.5v15m7.5-7.5h-15" />
+              </svg>
+              Add Preset
+            </button>
+          </div>
+          <p className="text-xs text-gray-500">Image generation presets with trigger conditions. Default trigger is "self-portrait".</p>
+
+          {(!character.gaze_presets || character.gaze_presets.length === 0) ? (
+            <p className="text-sm text-gray-600 italic">No GAZE presets configured.</p>
+          ) : (
+            <div className="space-y-3">
+              {character.gaze_presets.map((gp, idx) => (
+                <div key={idx} className="flex items-center gap-2 border border-gray-700 p-3 rounded-lg bg-gray-800/50">
+                  <div className="flex-1">
+                    <label className="block text-xs text-gray-500 mb-1">Preset</label>
+                    {isLoadingGaze ? (
+                      <p className="text-sm text-gray-500">Loading...</p>
+                    ) : availableGazePresets.length > 0 ? (
+                      <select
+                        className={selectClass + " text-sm"}
+                        value={gp.preset || ''}
+                        onChange={(e) => handleGazePresetChange(idx, 'preset', e.target.value)}
+                      >
+                        <option value="">-- Select --</option>
+                        {availableGazePresets.map(p => (
+                          <option key={p.slug} value={p.slug}>{p.name} ({p.slug})</option>
+                        ))}
+                      </select>
+                    ) : (
+                      <input
+                        type="text"
+                        className={inputClass + " text-sm"}
+                        value={gp.preset || ''}
+                        onChange={(e) => handleGazePresetChange(idx, 'preset', e.target.value)}
+                        placeholder="Preset slug"
+                      />
+                    )}
+                  </div>
+                  <div className="flex-1">
+                    <label className="block text-xs text-gray-500 mb-1">Trigger</label>
+                    <input
+                      type="text"
+                      className={inputClass + " text-sm"}
+                      value={gp.trigger || ''}
+                      onChange={(e) => handleGazePresetChange(idx, 'trigger', e.target.value)}
+                      placeholder="e.g. self-portrait, battle scene"
+                    />
+                  </div>
+                  <button
+                    onClick={() => removeGazePreset(idx)}
+                    className="mt-5 px-2 py-1.5 text-gray-500 hover:text-red-400 transition-colors"
+                    title="Remove"
+                  >
+                    <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+                      <path strokeLinecap="round" strokeLinejoin="round" d="M6 18L18 6M6 6l12 12" />
+                    </svg>
+                  </button>
+                </div>
+              ))}
             </div>
-          ))}
+          )}
         </div>
 
         {/* Model Overrides */}
@@ -509,7 +832,7 @@ export default function Editor() {
           <h2 className="text-lg font-semibold text-gray-200">Model Overrides</h2>
           <div>
             <label className={labelClass}>Primary Model</label>
-            <select className={selectClass} value={character.model_overrides?.primary || 'llama3.3:70b'} onChange={(e) => handleNestedChange('model_overrides', 'primary', e.target.value)}>
+            <select className={selectClass} value={character.model_overrides?.primary || 'qwen3.5:35b-a3b'} onChange={(e) => handleNestedChange('model_overrides', 'primary', e.target.value)}>
               <option value="llama3.3:70b">llama3.3:70b</option>
               <option value="qwen3.5:35b-a3b">qwen3.5:35b-a3b</option>
               <option value="qwen2.5:7b">qwen2.5:7b</option>
@@ -576,6 +899,17 @@ export default function Editor() {
           </div>
         )}
       </div>
+
+      {/* Notes */}
+      <div className={cardClass}>
+        <h2 className="text-lg font-semibold text-gray-200">Notes</h2>
+        <textarea
+          className={inputClass + " h-20 resize-y text-sm"}
+          value={character.notes || ''}
+          onChange={(e) => handleChange('notes', e.target.value)}
+          placeholder="Internal notes, reminders, or references..."
+        />
+      </div>
     </div>
   );
 }

homeai-dashboard/src/pages/Memories.jsx

@@ -0,0 +1,346 @@
+import { useState, useEffect, useCallback } from 'react';
+import {
+  getPersonalMemories, savePersonalMemory, deletePersonalMemory,
+  getGeneralMemories, saveGeneralMemory, deleteGeneralMemory,
+} from '../lib/memoryApi';
+
+const PERSONAL_CATEGORIES = [
+  { value: 'personal_info', label: 'Personal Info', color: 'bg-blue-500/20 text-blue-300 border-blue-500/30' },
+  { value: 'preference', label: 'Preference', color: 'bg-amber-500/20 text-amber-300 border-amber-500/30' },
+  { value: 'interaction', label: 'Interaction', color: 'bg-emerald-500/20 text-emerald-300 border-emerald-500/30' },
+  { value: 'emotional', label: 'Emotional', color: 'bg-pink-500/20 text-pink-300 border-pink-500/30' },
+  { value: 'other', label: 'Other', color: 'bg-gray-500/20 text-gray-300 border-gray-500/30' },
+];
+
+const GENERAL_CATEGORIES = [
+  { value: 'system', label: 'System', color: 'bg-indigo-500/20 text-indigo-300 border-indigo-500/30' },
+  { value: 'tool_usage', label: 'Tool Usage', color: 'bg-cyan-500/20 text-cyan-300 border-cyan-500/30' },
+  { value: 'home_layout', label: 'Home Layout', color: 'bg-emerald-500/20 text-emerald-300 border-emerald-500/30' },
+  { value: 'device', label: 'Device', color: 'bg-amber-500/20 text-amber-300 border-amber-500/30' },
+  { value: 'routine', label: 'Routine', color: 'bg-purple-500/20 text-purple-300 border-purple-500/30' },
+  { value: 'other', label: 'Other', color: 'bg-gray-500/20 text-gray-300 border-gray-500/30' },
+];
+
+const ACTIVE_KEY = 'homeai_active_character';
+
+function CategoryBadge({ category, categories }) {
+  const cat = categories.find(c => c.value === category) || categories[categories.length - 1];
+  return (
+    <span className={`px-2 py-0.5 text-xs rounded-full border ${cat.color}`}>
+      {cat.label}
+    </span>
+  );
+}
+
+function MemoryCard({ memory, categories, onEdit, onDelete }) {
+  return (
+    <div className="border border-gray-700 rounded-lg p-4 bg-gray-800/50 space-y-2">
+      <div className="flex items-start justify-between gap-3">
+        <p className="text-sm text-gray-200 flex-1 whitespace-pre-wrap">{memory.content}</p>
+        <div className="flex gap-1 shrink-0">
+          <button
+            onClick={() => onEdit(memory)}
+            className="p-1.5 text-gray-500 hover:text-gray-300 transition-colors"
+            title="Edit"
+          >
+            <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+              <path strokeLinecap="round" strokeLinejoin="round" d="M16.862 4.487l1.687-1.688a1.875 1.875 0 112.652 2.652L10.582 16.07a4.5 4.5 0 01-1.897 1.13L6 18l.8-2.685a4.5 4.5 0 011.13-1.897l8.932-8.931z" />
+            </svg>
+          </button>
+          <button
+            onClick={() => onDelete(memory.id)}
+            className="p-1.5 text-gray-500 hover:text-red-400 transition-colors"
+            title="Delete"
+          >
+            <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+              <path strokeLinecap="round" strokeLinejoin="round" d="M14.74 9l-.346 9m-4.788 0L9.26 9m9.968-3.21c.342.052.682.107 1.022.166m-1.022-.165L18.16 19.673a2.25 2.25 0 01-2.244 2.077H8.084a2.25 2.25 0 01-2.244-2.077L4.772 5.79m14.456 0a48.108 48.108 0 00-3.478-.397m-12 .562c.34-.059.68-.114 1.022-.165m0 0a48.11 48.11 0 013.478-.397m7.5 0v-.916c0-1.18-.91-2.164-2.09-2.201a51.964 51.964 0 00-3.32 0c-1.18.037-2.09 1.022-2.09 2.201v.916m7.5 0a48.667 48.667 0 00-7.5 0" />
+            </svg>
+          </button>
+        </div>
+      </div>
+      <div className="flex items-center gap-2">
+        <CategoryBadge category={memory.category} categories={categories} />
+        <span className="text-xs text-gray-600">
+          {memory.createdAt ? new Date(memory.createdAt).toLocaleDateString() : ''}
+        </span>
+      </div>
+    </div>
+  );
+}
+
+function MemoryForm({ categories, editing, onSave, onCancel }) {
+  const [content, setContent] = useState(editing?.content || '');
+  const [category, setCategory] = useState(editing?.category || categories[0].value);
+
+  const handleSubmit = () => {
+    if (!content.trim()) return;
+    const memory = {
+      ...(editing?.id ? { id: editing.id } : {}),
+      content: content.trim(),
+      category,
+    };
+    onSave(memory);
+    setContent('');
+    setCategory(categories[0].value);
+  };
+
+  return (
+    <div className="border border-indigo-500/30 rounded-lg p-4 bg-indigo-500/5 space-y-3">
+      <textarea
+        className="w-full bg-gray-800 border border-gray-700 text-gray-200 p-2 rounded-lg text-sm h-20 resize-y focus:border-indigo-500 focus:ring-1 focus:ring-indigo-500 outline-none"
+        value={content}
+        onChange={(e) => setContent(e.target.value)}
+        placeholder="Enter memory content..."
+        autoFocus
+      />
+      <div className="flex items-center gap-3">
+        <select
+          className="bg-gray-800 border border-gray-700 text-gray-200 text-sm p-2 rounded-lg focus:border-indigo-500 outline-none"
+          value={category}
+          onChange={(e) => setCategory(e.target.value)}
+        >
+          {categories.map(c => (
+            <option key={c.value} value={c.value}>{c.label}</option>
+          ))}
+        </select>
+        <div className="flex gap-2 ml-auto">
+          <button
+            onClick={onCancel}
+            className="px-3 py-1.5 bg-gray-700 hover:bg-gray-600 text-gray-300 text-sm rounded-lg transition-colors"
+          >
+            Cancel
+          </button>
+          <button
+            onClick={handleSubmit}
+            disabled={!content.trim()}
+            className="px-3 py-1.5 bg-indigo-600 hover:bg-indigo-500 disabled:bg-gray-700 disabled:text-gray-500 text-white text-sm rounded-lg transition-colors"
+          >
+            {editing?.id ? 'Update' : 'Add Memory'}
+          </button>
+        </div>
+      </div>
+    </div>
+  );
+}
+
+export default function Memories() {
+  const [tab, setTab] = useState('personal'); // 'personal' | 'general'
+  const [characters, setCharacters] = useState([]);
+  const [selectedCharId, setSelectedCharId] = useState('');
+  const [memories, setMemories] = useState([]);
+  const [loading, setLoading] = useState(false);
+  const [showForm, setShowForm] = useState(false);
+  const [editing, setEditing] = useState(null);
+  const [error, setError] = useState(null);
+  const [filter, setFilter] = useState('');
+
+  // Load characters list
+  useEffect(() => {
+    fetch('/api/characters')
+      .then(r => r.json())
+      .then(chars => {
+        setCharacters(chars);
+        const activeId = localStorage.getItem(ACTIVE_KEY);
+        if (activeId && chars.some(c => c.id === activeId)) {
+          setSelectedCharId(activeId);
+        } else if (chars.length > 0) {
+          setSelectedCharId(chars[0].id);
+        }
+      })
+      .catch(() => {});
+  }, []);
+
+  // Load memories when tab or selected character changes
+  const loadMemories = useCallback(async () => {
+    setLoading(true);
+    setError(null);
+    try {
+      if (tab === 'personal' && selectedCharId) {
+        const data = await getPersonalMemories(selectedCharId);
+        setMemories(data.memories || []);
+      } else if (tab === 'general') {
+        const data = await getGeneralMemories();
+        setMemories(data.memories || []);
+      } else {
+        setMemories([]);
+      }
+    } catch (err) {
+      setError(err.message);
+    } finally {
+      setLoading(false);
+    }
+  }, [tab, selectedCharId]);
+
+  useEffect(() => { loadMemories(); }, [loadMemories]);
+
+  const handleSave = async (memory) => {
+    try {
+      if (tab === 'personal') {
+        await savePersonalMemory(selectedCharId, memory);
+      } else {
+        await saveGeneralMemory(memory);
+      }
+      setShowForm(false);
+      setEditing(null);
+      await loadMemories();
+    } catch (err) {
+      setError(err.message);
+    }
+  };
+
+  const handleDelete = async (memoryId) => {
+    try {
+      if (tab === 'personal') {
+        await deletePersonalMemory(selectedCharId, memoryId);
+      } else {
+        await deleteGeneralMemory(memoryId);
+      }
+      await loadMemories();
+    } catch (err) {
+      setError(err.message);
+    }
+  };
+
+  const handleEdit = (memory) => {
+    setEditing(memory);
+    setShowForm(true);
+  };
+
+  const categories = tab === 'personal' ? PERSONAL_CATEGORIES : GENERAL_CATEGORIES;
+  const filteredMemories = filter
+    ? memories.filter(m => m.content?.toLowerCase().includes(filter.toLowerCase()) || m.category === filter)
+    : memories;
+
+  // Sort newest first
+  const sortedMemories = [...filteredMemories].sort(
+    (a, b) => (b.createdAt || '').localeCompare(a.createdAt || '')
+  );
+
+  const selectedChar = characters.find(c => c.id === selectedCharId);
+
+  return (
+    <div className="space-y-6">
+      {/* Header */}
+      <div className="flex items-center justify-between">
+        <div>
+          <h1 className="text-3xl font-bold text-gray-100">Memories</h1>
+          <p className="text-sm text-gray-500 mt-1">
+            {sortedMemories.length} {tab} memor{sortedMemories.length !== 1 ? 'ies' : 'y'}
+            {tab === 'personal' && selectedChar && (
+              <span className="ml-1 text-indigo-400">
+                for {selectedChar.data?.display_name || selectedChar.data?.name || selectedCharId}
+              </span>
+            )}
+          </p>
+        </div>
+        <button
+          onClick={() => { setEditing(null); setShowForm(!showForm); }}
+          className="flex items-center gap-2 px-4 py-2 bg-indigo-600 hover:bg-indigo-500 text-white rounded-lg transition-colors"
+        >
+          <svg className="w-4 h-4" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={2}>
+            <path strokeLinecap="round" strokeLinejoin="round" d="M12 4.5v15m7.5-7.5h-15" />
+          </svg>
+          Add Memory
+        </button>
+      </div>
+
+      {error && (
+        <div className="bg-red-900/30 border border-red-500/50 text-red-300 px-4 py-3 rounded-lg text-sm">
+          {error}
+          <button onClick={() => setError(null)} className="ml-2 text-red-400 hover:text-red-300">&times;</button>
+        </div>
+      )}
+
+      {/* Tabs */}
+      <div className="flex gap-1 bg-gray-900 p-1 rounded-lg border border-gray-800 w-fit">
+        <button
+          onClick={() => { setTab('personal'); setShowForm(false); setEditing(null); }}
+          className={`px-4 py-2 text-sm font-medium rounded-md transition-colors ${
+            tab === 'personal'
+              ? 'bg-gray-800 text-white'
+              : 'text-gray-400 hover:text-gray-200'
+          }`}
+        >
+          Personal
+        </button>
+        <button
+          onClick={() => { setTab('general'); setShowForm(false); setEditing(null); }}
+          className={`px-4 py-2 text-sm font-medium rounded-md transition-colors ${
+            tab === 'general'
+              ? 'bg-gray-800 text-white'
+              : 'text-gray-400 hover:text-gray-200'
+          }`}
+        >
+          General
+        </button>
+      </div>
+
+      {/* Character selector (personal tab only) */}
+      {tab === 'personal' && (
+        <div className="flex items-center gap-3">
+          <label className="text-sm text-gray-400">Character</label>
+          <select
+            value={selectedCharId}
+            onChange={(e) => setSelectedCharId(e.target.value)}
+            className="bg-gray-800 border border-gray-700 text-gray-200 text-sm p-2 rounded-lg focus:border-indigo-500 outline-none"
+          >
+            {characters.map(c => (
+              <option key={c.id} value={c.id}>
+                {c.data?.display_name || c.data?.name || c.id}
+              </option>
+            ))}
+          </select>
+        </div>
+      )}
+
+      {/* Search filter */}
+      <div>
+        <input
+          type="text"
+          className="w-full bg-gray-800 border border-gray-700 text-gray-200 p-2 rounded-lg text-sm focus:border-indigo-500 focus:ring-1 focus:ring-indigo-500 outline-none"
+          value={filter}
+          onChange={(e) => setFilter(e.target.value)}
+          placeholder="Search memories..."
+        />
+      </div>
+
+      {/* Add/Edit form */}
+      {showForm && (
+        <MemoryForm
+          categories={categories}
+          editing={editing}
+          onSave={handleSave}
+          onCancel={() => { setShowForm(false); setEditing(null); }}
+        />
+      )}
+
+      {/* Memory list */}
+      {loading ? (
+        <div className="text-center py-12">
+          <p className="text-gray-500">Loading memories...</p>
+        </div>
+      ) : sortedMemories.length === 0 ? (
+        <div className="text-center py-12">
+          <svg className="w-12 h-12 mx-auto text-gray-700 mb-3" fill="none" viewBox="0 0 24 24" stroke="currentColor" strokeWidth={1}>
+            <path strokeLinecap="round" strokeLinejoin="round" d="M12 18v-5.25m0 0a6.01 6.01 0 001.5-.189m-1.5.189a6.01 6.01 0 01-1.5-.189m3.75 7.478a12.06 12.06 0 01-4.5 0m3.75 2.383a14.406 14.406 0 01-3 0M14.25 18v-.192c0-.983.658-1.823 1.508-2.316a7.5 7.5 0 10-7.517 0c.85.493 1.509 1.333 1.509 2.316V18" />
+          </svg>
+          <p className="text-gray-500 text-sm">
+            {filter ? 'No memories match your search.' : 'No memories yet. Add one to get started.'}
+          </p>
+        </div>
+      ) : (
+        <div className="space-y-3">
+          {sortedMemories.map(memory => (
+            <MemoryCard
+              key={memory.id}
+              memory={memory}
+              categories={categories}
+              onEdit={handleEdit}
+              onDelete={handleDelete}
+            />
+          ))}
+        </div>
+      )}
+    </div>
+  );
+}

homeai-dashboard/vite.config.js

@@ -2,6 +2,267 @@ import { defineConfig } from 'vite'
 import react from '@vitejs/plugin-react'
 import tailwindcss from '@tailwindcss/vite'
 
+const CHARACTERS_DIR = '/Users/aodhan/homeai-data/characters'
+const SATELLITE_MAP_PATH = '/Users/aodhan/homeai-data/satellite-map.json'
+const CONVERSATIONS_DIR = '/Users/aodhan/homeai-data/conversations'
+const MEMORIES_DIR = '/Users/aodhan/homeai-data/memories'
+const GAZE_HOST = 'http://10.0.0.101:5782'
+const GAZE_API_KEY = process.env.GAZE_API_KEY || ''
+
+function characterStoragePlugin() {
+  return {
+    name: 'character-storage',
+    configureServer(server) {
+      const ensureDir = async () => {
+        const { mkdir } = await import('fs/promises')
+        await mkdir(CHARACTERS_DIR, { recursive: true })
+      }
+
+      // GET /api/characters — list all profiles
+      server.middlewares.use('/api/characters', async (req, res, next) => {
+        if (req.method === 'OPTIONS') {
+          res.writeHead(204, { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'GET,POST,DELETE', 'Access-Control-Allow-Headers': 'Content-Type' })
+          res.end()
+          return
+        }
+
+        const { readdir, readFile, writeFile, unlink } = await import('fs/promises')
+        await ensureDir()
+
+        // req.url has the mount prefix stripped by connect, so "/" means /api/characters
+        const url = new URL(req.url, 'http://localhost')
+        const subPath = url.pathname.replace(/^\/+/, '')
+
+        // GET /api/characters/:id — single profile
+        if (req.method === 'GET' && subPath) {
+          try {
+            const safeId = subPath.replace(/[^a-zA-Z0-9_\-\.]/g, '_')
+            const raw = await readFile(`${CHARACTERS_DIR}/${safeId}.json`, 'utf-8')
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(raw)
+          } catch {
+            res.writeHead(404, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ error: 'Not found' }))
+          }
+          return
+        }
+
+        if (req.method === 'GET' && !subPath) {
+          try {
+            const files = (await readdir(CHARACTERS_DIR)).filter(f => f.endsWith('.json'))
+            const profiles = []
+            for (const file of files) {
+              try {
+                const raw = await readFile(`${CHARACTERS_DIR}/${file}`, 'utf-8')
+                profiles.push(JSON.parse(raw))
+              } catch { /* skip corrupt files */ }
+            }
+            // Sort by addedAt descending
+            profiles.sort((a, b) => (b.addedAt || '').localeCompare(a.addedAt || ''))
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify(profiles))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        if (req.method === 'POST' && !subPath) {
+          try {
+            const chunks = []
+            for await (const chunk of req) chunks.push(chunk)
+            const profile = JSON.parse(Buffer.concat(chunks).toString())
+            if (!profile.id) {
+              res.writeHead(400, { 'Content-Type': 'application/json' })
+              res.end(JSON.stringify({ error: 'Missing profile id' }))
+              return
+            }
+            // Sanitize filename — only allow alphanumeric, underscore, dash, dot
+            const safeId = profile.id.replace(/[^a-zA-Z0-9_\-\.]/g, '_')
+            await writeFile(`${CHARACTERS_DIR}/${safeId}.json`, JSON.stringify(profile, null, 2))
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        if (req.method === 'DELETE' && subPath) {
+          try {
+            const safeId = subPath.replace(/[^a-zA-Z0-9_\-\.]/g, '_')
+            await unlink(`${CHARACTERS_DIR}/${safeId}.json`).catch(() => {})
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        next()
+      })
+    },
+  }
+}
+
+function satelliteMapPlugin() {
+  return {
+    name: 'satellite-map',
+    configureServer(server) {
+      server.middlewares.use('/api/satellite-map', async (req, res, next) => {
+        if (req.method === 'OPTIONS') {
+          res.writeHead(204, { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'GET,POST', 'Access-Control-Allow-Headers': 'Content-Type' })
+          res.end()
+          return
+        }
+
+        const { readFile, writeFile } = await import('fs/promises')
+
+        if (req.method === 'GET') {
+          try {
+            const raw = await readFile(SATELLITE_MAP_PATH, 'utf-8')
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(raw)
+          } catch {
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ default: 'aria_default', satellites: {} }))
+          }
+          return
+        }
+
+        if (req.method === 'POST') {
+          try {
+            const chunks = []
+            for await (const chunk of req) chunks.push(chunk)
+            const data = JSON.parse(Buffer.concat(chunks).toString())
+            await writeFile(SATELLITE_MAP_PATH, JSON.stringify(data, null, 2))
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        next()
+      })
+    },
+  }
+}
+
+function conversationStoragePlugin() {
+  return {
+    name: 'conversation-storage',
+    configureServer(server) {
+      const ensureDir = async () => {
+        const { mkdir } = await import('fs/promises')
+        await mkdir(CONVERSATIONS_DIR, { recursive: true })
+      }
+
+      server.middlewares.use('/api/conversations', async (req, res, next) => {
+        if (req.method === 'OPTIONS') {
+          res.writeHead(204, { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'GET,POST,DELETE', 'Access-Control-Allow-Headers': 'Content-Type' })
+          res.end()
+          return
+        }
+
+        const { readdir, readFile, writeFile, unlink } = await import('fs/promises')
+        await ensureDir()
+
+        const url = new URL(req.url, 'http://localhost')
+        const subPath = url.pathname.replace(/^\/+/, '')
+
+        // GET /api/conversations/:id — single conversation with messages
+        if (req.method === 'GET' && subPath) {
+          try {
+            const safeId = subPath.replace(/[^a-zA-Z0-9_\-\.]/g, '_')
+            const raw = await readFile(`${CONVERSATIONS_DIR}/${safeId}.json`, 'utf-8')
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(raw)
+          } catch {
+            res.writeHead(404, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ error: 'Not found' }))
+          }
+          return
+        }
+
+        // GET /api/conversations — list metadata (no messages)
+        if (req.method === 'GET' && !subPath) {
+          try {
+            const files = (await readdir(CONVERSATIONS_DIR)).filter(f => f.endsWith('.json'))
+            const list = []
+            for (const file of files) {
+              try {
+                const raw = await readFile(`${CONVERSATIONS_DIR}/${file}`, 'utf-8')
+                const conv = JSON.parse(raw)
+                list.push({
+                  id: conv.id,
+                  title: conv.title || '',
+                  characterId: conv.characterId || '',
+                  characterName: conv.characterName || '',
+                  createdAt: conv.createdAt || '',
+                  updatedAt: conv.updatedAt || '',
+                  messageCount: (conv.messages || []).length,
+                })
+              } catch { /* skip corrupt files */ }
+            }
+            list.sort((a, b) => (b.updatedAt || '').localeCompare(a.updatedAt || ''))
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify(list))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        // POST /api/conversations — create or update
+        if (req.method === 'POST' && !subPath) {
+          try {
+            const chunks = []
+            for await (const chunk of req) chunks.push(chunk)
+            const conv = JSON.parse(Buffer.concat(chunks).toString())
+            if (!conv.id) {
+              res.writeHead(400, { 'Content-Type': 'application/json' })
+              res.end(JSON.stringify({ error: 'Missing conversation id' }))
+              return
+            }
+            const safeId = conv.id.replace(/[^a-zA-Z0-9_\-\.]/g, '_')
+            await writeFile(`${CONVERSATIONS_DIR}/${safeId}.json`, JSON.stringify(conv, null, 2))
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        // DELETE /api/conversations/:id
+        if (req.method === 'DELETE' && subPath) {
+          try {
+            const safeId = subPath.replace(/[^a-zA-Z0-9_\-\.]/g, '_')
+            await unlink(`${CONVERSATIONS_DIR}/${safeId}.json`).catch(() => {})
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        next()
+      })
+    },
+  }
+}
+
 function healthCheckPlugin() {
   return {
     name: 'health-check-proxy',
@@ -121,6 +382,273 @@ function healthCheckPlugin() {
   };
 }
 
+function gazeProxyPlugin() {
+  return {
+    name: 'gaze-proxy',
+    configureServer(server) {
+      server.middlewares.use('/api/gaze/presets', async (req, res) => {
+        if (!GAZE_API_KEY) {
+          res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(JSON.stringify({ presets: [] }))
+          return
+        }
+        try {
+          const http = await import('http')
+          const url = new URL(`${GAZE_HOST}/api/v1/presets`)
+          const proxyRes = await new Promise((resolve, reject) => {
+            const r = http.default.get(url, { headers: { 'X-API-Key': GAZE_API_KEY }, timeout: 5000 }, resolve)
+            r.on('error', reject)
+            r.on('timeout', () => { r.destroy(); reject(new Error('timeout')) })
+          })
+          const chunks = []
+          for await (const chunk of proxyRes) chunks.push(chunk)
+          res.writeHead(proxyRes.statusCode, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(Buffer.concat(chunks))
+        } catch {
+          res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(JSON.stringify({ presets: [] }))
+        }
+      })
+    },
+  }
+}
+
+function memoryStoragePlugin() {
+  return {
+    name: 'memory-storage',
+    configureServer(server) {
+      const ensureDirs = async () => {
+        const { mkdir } = await import('fs/promises')
+        await mkdir(`${MEMORIES_DIR}/personal`, { recursive: true })
+      }
+
+      const readJsonFile = async (path, fallback) => {
+        const { readFile } = await import('fs/promises')
+        try {
+          return JSON.parse(await readFile(path, 'utf-8'))
+        } catch {
+          return fallback
+        }
+      }
+
+      const writeJsonFile = async (path, data) => {
+        const { writeFile } = await import('fs/promises')
+        await writeFile(path, JSON.stringify(data, null, 2))
+      }
+
+      // Personal memories: /api/memories/personal/:characterId[/:memoryId]
+      server.middlewares.use('/api/memories/personal', async (req, res, next) => {
+        if (req.method === 'OPTIONS') {
+          res.writeHead(204, { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'GET,POST,DELETE', 'Access-Control-Allow-Headers': 'Content-Type' })
+          res.end()
+          return
+        }
+
+        await ensureDirs()
+        const url = new URL(req.url, 'http://localhost')
+        const parts = url.pathname.replace(/^\/+/, '').split('/')
+        const characterId = parts[0] ? parts[0].replace(/[^a-zA-Z0-9_\-\.]/g, '_') : null
+        const memoryId = parts[1] || null
+
+        if (!characterId) {
+          res.writeHead(400, { 'Content-Type': 'application/json' })
+          res.end(JSON.stringify({ error: 'Missing character ID' }))
+          return
+        }
+
+        const filePath = `${MEMORIES_DIR}/personal/${characterId}.json`
+
+        if (req.method === 'GET') {
+          const data = await readJsonFile(filePath, { characterId, memories: [] })
+          res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(JSON.stringify(data))
+          return
+        }
+
+        if (req.method === 'POST') {
+          try {
+            const chunks = []
+            for await (const chunk of req) chunks.push(chunk)
+            const memory = JSON.parse(Buffer.concat(chunks).toString())
+            const data = await readJsonFile(filePath, { characterId, memories: [] })
+            if (memory.id) {
+              const idx = data.memories.findIndex(m => m.id === memory.id)
+              if (idx >= 0) {
+                data.memories[idx] = { ...data.memories[idx], ...memory }
+              } else {
+                data.memories.push(memory)
+              }
+            } else {
+              memory.id = 'm_' + Date.now()
+              memory.createdAt = memory.createdAt || new Date().toISOString()
+              data.memories.push(memory)
+            }
+            await writeJsonFile(filePath, data)
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true, memory }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        if (req.method === 'DELETE' && memoryId) {
+          try {
+            const data = await readJsonFile(filePath, { characterId, memories: [] })
+            data.memories = data.memories.filter(m => m.id !== memoryId)
+            await writeJsonFile(filePath, data)
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        next()
+      })
+
+      // General memories: /api/memories/general[/:memoryId]
+      server.middlewares.use('/api/memories/general', async (req, res, next) => {
+        if (req.method === 'OPTIONS') {
+          res.writeHead(204, { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'GET,POST,DELETE', 'Access-Control-Allow-Headers': 'Content-Type' })
+          res.end()
+          return
+        }
+
+        await ensureDirs()
+        const url = new URL(req.url, 'http://localhost')
+        const memoryId = url.pathname.replace(/^\/+/, '') || null
+        const filePath = `${MEMORIES_DIR}/general.json`
+
+        if (req.method === 'GET') {
+          const data = await readJsonFile(filePath, { memories: [] })
+          res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(JSON.stringify(data))
+          return
+        }
+
+        if (req.method === 'POST') {
+          try {
+            const chunks = []
+            for await (const chunk of req) chunks.push(chunk)
+            const memory = JSON.parse(Buffer.concat(chunks).toString())
+            const data = await readJsonFile(filePath, { memories: [] })
+            if (memory.id) {
+              const idx = data.memories.findIndex(m => m.id === memory.id)
+              if (idx >= 0) {
+                data.memories[idx] = { ...data.memories[idx], ...memory }
+              } else {
+                data.memories.push(memory)
+              }
+            } else {
+              memory.id = 'm_' + Date.now()
+              memory.createdAt = memory.createdAt || new Date().toISOString()
+              data.memories.push(memory)
+            }
+            await writeJsonFile(filePath, data)
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true, memory }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        if (req.method === 'DELETE' && memoryId) {
+          try {
+            const data = await readJsonFile(filePath, { memories: [] })
+            data.memories = data.memories.filter(m => m.id !== memoryId)
+            await writeJsonFile(filePath, data)
+            res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+            res.end(JSON.stringify({ ok: true }))
+          } catch (err) {
+            res.writeHead(500, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: err.message }))
+          }
+          return
+        }
+
+        next()
+      })
+    },
+  }
+}
+
+function characterLookupPlugin() {
+  return {
+    name: 'character-lookup',
+    configureServer(server) {
+      server.middlewares.use('/api/character-lookup', async (req, res) => {
+        if (req.method === 'OPTIONS') {
+          res.writeHead(204, { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'POST', 'Access-Control-Allow-Headers': 'Content-Type' })
+          res.end()
+          return
+        }
+        if (req.method !== 'POST') {
+          res.writeHead(405, { 'Content-Type': 'application/json' })
+          res.end(JSON.stringify({ error: 'POST only' }))
+          return
+        }
+
+        try {
+          const chunks = []
+          for await (const chunk of req) chunks.push(chunk)
+          const { name, franchise } = JSON.parse(Buffer.concat(chunks).toString())
+          if (!name || !franchise) {
+            res.writeHead(400, { 'Content-Type': 'application/json' })
+            res.end(JSON.stringify({ error: 'Missing name or franchise' }))
+            return
+          }
+
+          const { execFile } = await import('child_process')
+          const { promisify } = await import('util')
+          const execFileAsync = promisify(execFile)
+
+          // Call the MCP fetcher inside the running Docker container
+          const safeName = name.replace(/'/g, "\\'")
+          const safeFranchise = franchise.replace(/'/g, "\\'")
+          const pyScript = `
+import asyncio, json
+from character_details.fetcher import fetch_character
+c = asyncio.run(fetch_character('${safeName}', '${safeFranchise}'))
+print(json.dumps(c.model_dump(), default=str))
+`.trim()
+
+          const { stdout } = await execFileAsync(
+            'docker',
+            ['exec', 'character-browser-character-mcp-1', 'python', '-c', pyScript],
+            { timeout: 30000 }
+          )
+
+          const data = JSON.parse(stdout.trim())
+          res.writeHead(200, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(JSON.stringify({
+            name: data.name || name,
+            franchise: data.franchise || franchise,
+            description: data.description || '',
+            background: data.background || '',
+            appearance: data.appearance || '',
+            personality: data.personality || '',
+            abilities: data.abilities || [],
+            notable_quotes: data.notable_quotes || [],
+            relationships: data.relationships || [],
+            sources: data.sources || [],
+          }))
+        } catch (err) {
+          console.error('[character-lookup] failed:', err?.message || err)
+          const status = err?.message?.includes('timeout') ? 504 : 500
+          res.writeHead(status, { 'Content-Type': 'application/json', 'Access-Control-Allow-Origin': '*' })
+          res.end(JSON.stringify({ error: err?.message || 'Lookup failed' }))
+        }
+      })
+    },
+  }
+}
+
 function bridgeProxyPlugin() {
   return {
     name: 'bridge-proxy',
@@ -172,10 +700,11 @@ function bridgeProxyPlugin() {
             proxyReq.write(body)
             proxyReq.end()
           })
-        } catch {
+        } catch (err) {
+          console.error(`[bridge-proxy] ${targetPath} failed:`, err?.message || err)
           if (!res.headersSent) {
             res.writeHead(502, { 'Content-Type': 'application/json' })
-            res.end(JSON.stringify({ error: 'Bridge unreachable' }))
+            res.end(JSON.stringify({ error: `Bridge unreachable: ${err?.message || 'unknown'}` }))
           }
         }
       }
@@ -189,6 +718,12 @@ function bridgeProxyPlugin() {
 
 export default defineConfig({
   plugins: [
+    characterStoragePlugin(),
+    satelliteMapPlugin(),
+    conversationStoragePlugin(),
+    memoryStoragePlugin(),
+    gazeProxyPlugin(),
+    characterLookupPlugin(),
     healthCheckPlugin(),
     bridgeProxyPlugin(),
     tailwindcss(),

homeai-images/API_GUIDE.md

@@ -0,0 +1,219 @@
+# GAZE REST API Guide
+
+## Setup
+
+1. Open **Settings** in the GAZE web UI
+2. Scroll to **REST API Key** and click **Regenerate**
+3. Copy the key — you'll need it for all API requests
+
+## Authentication
+
+Every request must include your API key via one of:
+
+- **Header (recommended):** `X-API-Key: <your-key>`
+- **Query parameter:** `?api_key=<your-key>`
+
+Responses for auth failures:
+
+| Status | Meaning |
+|--------|---------|
+| `401` | Missing API key |
+| `403` | Invalid API key |
+
+## Endpoints
+
+### List Presets
+
+```
+GET /api/v1/presets
+```
+
+Returns all available presets.
+
+**Response:**
+
+```json
+{
+  "presets": [
+    {
+      "preset_id": "example_01",
+      "slug": "example_01",
+      "name": "Example Preset",
+      "has_cover": true
+    }
+  ]
+}
+```
+
+### Generate Image
+
+```
+POST /api/v1/generate/<preset_slug>
+```
+
+Queue one or more image generations using a preset's configuration. All body parameters are optional — when omitted, the preset's own settings are used.
+
+**Request body (JSON):**
+
+| Field | Type | Default | Description |
+|-------|------|---------|-------------|
+| `count` | int | `1` | Number of images to generate (1–20) |
+| `checkpoint` | string | — | Override checkpoint path (e.g. `"Illustrious/model.safetensors"`) |
+| `extra_positive` | string | `""` | Additional positive prompt tags appended to the generated prompt |
+| `extra_negative` | string | `""` | Additional negative prompt tags |
+| `seed` | int | random | Fixed seed for reproducible generation |
+| `width` | int | — | Output width in pixels (must provide both width and height) |
+| `height` | int | — | Output height in pixels (must provide both width and height) |
+
+**Response (202):**
+
+```json
+{
+  "jobs": [
+    { "job_id": "783f0268-ba85-4426-8ca2-6393c844c887", "status": "queued" }
+  ]
+}
+```
+
+**Errors:**
+
+| Status | Cause |
+|--------|-------|
+| `400` | Invalid parameters (bad count, seed, or mismatched width/height) |
+| `404` | Preset slug not found |
+| `500` | Internal generation error |
+
+### Check Job Status
+
+```
+GET /api/v1/job/<job_id>
+```
+
+Poll this endpoint to track generation progress.
+
+**Response:**
+
+```json
+{
+  "id": "783f0268-ba85-4426-8ca2-6393c844c887",
+  "label": "Preset: Example Preset – preview",
+  "status": "done",
+  "error": null,
+  "result": {
+    "image_url": "/static/uploads/presets/example_01/gen_1773601346.png",
+    "relative_path": "presets/example_01/gen_1773601346.png",
+    "seed": 927640517599332
+  }
+}
+```
+
+**Job statuses:**
+
+| Status | Meaning |
+|--------|---------|
+| `pending` | Waiting in queue |
+| `processing` | Currently generating |
+| `done` | Complete — `result` contains image info |
+| `failed` | Error occurred — check `error` field |
+
+The `result` object is only present when status is `done`. Use `seed` from the result to reproduce the exact same image later.
+
+**Retrieving the image:** The `image_url` is a path relative to the server root. Fetch it directly:
+
+```
+GET http://<host>:5782/static/uploads/presets/example_01/gen_1773601346.png
+```
+
+Image retrieval does not require authentication.
+
+## Examples
+
+### Generate a single image and wait for it
+
+```bash
+API_KEY="your-key-here"
+HOST="http://localhost:5782"
+
+# Queue generation
+JOB_ID=$(curl -s -X POST \
+  -H "X-API-Key: $API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{}' \
+  "$HOST/api/v1/generate/example_01" | python3 -c "import sys,json; print(json.load(sys.stdin)['jobs'][0]['job_id'])")
+
+echo "Job: $JOB_ID"
+
+# Poll until done
+while true; do
+  RESULT=$(curl -s -H "X-API-Key: $API_KEY" "$HOST/api/v1/job/$JOB_ID")
+  STATUS=$(echo "$RESULT" | python3 -c "import sys,json; print(json.load(sys.stdin)['status'])")
+  echo "Status: $STATUS"
+  if [ "$STATUS" = "done" ] || [ "$STATUS" = "failed" ]; then
+    echo "$RESULT" | python3 -m json.tool
+    break
+  fi
+  sleep 5
+done
+```
+
+### Generate 3 images with extra prompts
+
+```bash
+curl -X POST \
+  -H "X-API-Key: $API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "count": 3,
+    "extra_positive": "smiling, outdoors",
+    "extra_negative": "blurry"
+  }' \
+  "$HOST/api/v1/generate/example_01"
+```
+
+### Reproduce a specific image
+
+```bash
+curl -X POST \
+  -H "X-API-Key: $API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{"seed": 927640517599332}' \
+  "$HOST/api/v1/generate/example_01"
+```
+
+### Python example
+
+```python
+import requests
+import time
+
+HOST = "http://localhost:5782"
+API_KEY = "your-key-here"
+HEADERS = {"X-API-Key": API_KEY, "Content-Type": "application/json"}
+
+# List presets
+presets = requests.get(f"{HOST}/api/v1/presets", headers=HEADERS).json()
+print(f"Available presets: {[p['name'] for p in presets['presets']]}")
+
+# Generate
+resp = requests.post(
+    f"{HOST}/api/v1/generate/{presets['presets'][0]['slug']}",
+    headers=HEADERS,
+    json={"count": 1},
+).json()
+
+job_id = resp["jobs"][0]["job_id"]
+print(f"Queued job: {job_id}")
+
+# Poll
+while True:
+    status = requests.get(f"{HOST}/api/v1/job/{job_id}", headers=HEADERS).json()
+    print(f"Status: {status['status']}")
+    if status["status"] in ("done", "failed"):
+        break
+    time.sleep(5)
+
+if status["status"] == "done":
+    image_url = f"{HOST}{status['result']['image_url']}"
+    print(f"Image: {image_url}")
+    print(f"Seed: {status['result']['seed']}")
+```

homeai-llm/launchd/com.homeai.preload-models.plist

@@ -12,17 +12,27 @@
     <string>/Users/aodhan/gitea/homeai/homeai-llm/scripts/preload-models.sh</string>
   </array>
 
+  <key>EnvironmentVariables</key>
+  <dict>
+    <!-- Override to change which medium model stays warm -->
+    <key>HOMEAI_MEDIUM_MODEL</key>
+    <string>qwen3.5:35b-a3b</string>
+  </dict>
+
   <key>RunAtLoad</key>
   <true/>
 
+  <key>KeepAlive</key>
+  <true/>
+
   <key>StandardOutPath</key>
   <string>/tmp/homeai-preload-models.log</string>
 
   <key>StandardErrorPath</key>
   <string>/tmp/homeai-preload-models-error.log</string>
 
-  <!-- Delay 15s to let Ollama start first -->
+  <!-- If the script exits/crashes, wait 30s before restarting -->
   <key>ThrottleInterval</key>
-  <integer>15</integer>
+  <integer>30</integer>
 </dict>
 </plist>

homeai-llm/scripts/preload-models.sh

@@ -1,19 +1,73 @@
 #!/bin/bash
-# Pre-load voice pipeline models into Ollama with infinite keep_alive.
+# Keep voice pipeline models warm in Ollama VRAM.
-# Run after Ollama starts (called by launchd or manually).
+# Runs as a loop — checks every 5 minutes, re-pins any model that got evicted.
 # Only pins lightweight/MoE models — large dense models (70B) use default expiry.
 
 OLLAMA_URL="http://localhost:11434"
+CHECK_INTERVAL=300  # seconds between checks
 
-# Wait for Ollama to be ready
+# Medium model can be overridden via env var (e.g. by persona config)
-for i in $(seq 1 30); do
+HOMEAI_MEDIUM_MODEL="${HOMEAI_MEDIUM_MODEL:-qwen3.5:35b-a3b}"
-    curl -sf "$OLLAMA_URL/api/tags" > /dev/null 2>&1 && break
+
-    sleep 2
+# Models to keep warm: "name|description"
+MODELS=(
+    "qwen2.5:7b|small (4.7GB) — fast fallback"
+    "${HOMEAI_MEDIUM_MODEL}|medium — persona default"
+)
+
+wait_for_ollama() {
+    for i in $(seq 1 30); do
+        curl -sf "$OLLAMA_URL/api/tags" > /dev/null 2>&1 && return 0
+        sleep 2
+    done
+    return 1
+}
+
+is_model_loaded() {
+    local model="$1"
+    curl -sf "$OLLAMA_URL/api/ps" 2>/dev/null \
+        | python3 -c "
+import json, sys
+data = json.load(sys.stdin)
+names = [m['name'] for m in data.get('models', [])]
+sys.exit(0 if '$model' in names else 1)
+" 2>/dev/null
+}
+
+pin_model() {
+    local model="$1"
+    local desc="$2"
+    if is_model_loaded "$model"; then
+        echo "[keepwarm] $model already loaded — skipping"
+        return 0
+    fi
+    echo "[keepwarm] Loading $model ($desc) with keep_alive=-1..."
+    curl -sf "$OLLAMA_URL/api/generate" \
+        -d "{\"model\":\"$model\",\"prompt\":\"ready\",\"stream\":false,\"keep_alive\":-1,\"options\":{\"num_ctx\":512}}" \
+        > /dev/null 2>&1
+    if [ $? -eq 0 ]; then
+        echo "[keepwarm] $model pinned in VRAM"
+    else
+        echo "[keepwarm] ERROR: failed to load $model"
+    fi
+}
+
+# --- Main loop ---
+
+echo "[keepwarm] Starting model keep-warm daemon (interval: ${CHECK_INTERVAL}s)"
+
+# Initial wait for Ollama
+if ! wait_for_ollama; then
+    echo "[keepwarm] ERROR: Ollama not reachable after 60s, exiting"
+    exit 1
+fi
+echo "[keepwarm] Ollama is online"
+
+while true; do
+    for entry in "${MODELS[@]}"; do
+        IFS='|' read -r model desc <<< "$entry"
+        pin_model "$model" "$desc"
+    done
+
+    sleep "$CHECK_INTERVAL"
 done
-
-# Pin qwen3.5:35b-a3b (MoE, 38.7GB VRAM, voice pipeline default)
-echo "[preload] Loading qwen3.5:35b-a3b with keep_alive=-1..."
-curl -sf "$OLLAMA_URL/api/generate" \
-    -d '{"model":"qwen3.5:35b-a3b","prompt":"ready","stream":false,"keep_alive":-1,"options":{"num_ctx":512}}' \
-    > /dev/null 2>&1
-echo "[preload] qwen3.5:35b-a3b pinned in memory"

homeai-visual/launchd/com.homeai.vtube-bridge.plist

@@ -0,0 +1,40 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN"
+  "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+  <key>Label</key>
+  <string>com.homeai.vtube-bridge</string>
+
+  <key>ProgramArguments</key>
+  <array>
+    <string>/Users/aodhan/homeai-visual-env/bin/python3</string>
+    <string>/Users/aodhan/gitea/homeai/homeai-visual/vtube-bridge.py</string>
+    <string>--port</string>
+    <string>8002</string>
+    <string>--character</string>
+    <string>/Users/aodhan/gitea/homeai/homeai-dashboard/characters/aria.json</string>
+  </array>
+
+  <key>RunAtLoad</key>
+  <true/>
+
+  <key>KeepAlive</key>
+  <true/>
+
+  <key>StandardOutPath</key>
+  <string>/tmp/homeai-vtube-bridge.log</string>
+
+  <key>StandardErrorPath</key>
+  <string>/tmp/homeai-vtube-bridge-error.log</string>
+
+  <key>ThrottleInterval</key>
+  <integer>10</integer>
+
+  <key>EnvironmentVariables</key>
+  <dict>
+    <key>PATH</key>
+    <string>/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin</string>
+  </dict>
+</dict>
+</plist>

homeai-visual/scripts/test-expressions.py

@@ -0,0 +1,170 @@
+#!/usr/bin/env python3
+"""
+Test script for VTube Studio Expression Bridge.
+
+Usage:
+    python3 test-expressions.py              # test all expressions
+    python3 test-expressions.py --auth       # run auth flow first
+    python3 test-expressions.py --lipsync    # test lip sync parameter
+    python3 test-expressions.py --latency    # measure round-trip latency
+
+Requires the vtube-bridge to be running on port 8002.
+"""
+
+import argparse
+import json
+import sys
+import time
+import urllib.request
+
+BRIDGE_URL = "http://localhost:8002"
+
+EXPRESSIONS = ["idle", "listening", "thinking", "speaking", "happy", "sad", "surprised", "error"]
+
+
+def _post(path: str, data: dict | None = None) -> dict:
+    body = json.dumps(data or {}).encode()
+    req = urllib.request.Request(
+        f"{BRIDGE_URL}{path}",
+        data=body,
+        headers={"Content-Type": "application/json"},
+        method="POST",
+    )
+    with urllib.request.urlopen(req, timeout=10) as resp:
+        return json.loads(resp.read())
+
+
+def _get(path: str) -> dict:
+    req = urllib.request.Request(f"{BRIDGE_URL}{path}")
+    with urllib.request.urlopen(req, timeout=10) as resp:
+        return json.loads(resp.read())
+
+
+def check_bridge():
+    """Verify bridge is running and connected."""
+    try:
+        status = _get("/status")
+        print(f"Bridge status: connected={status['connected']}, authenticated={status['authenticated']}")
+        print(f"  Expressions: {', '.join(status.get('expressions', []))}")
+        if not status["connected"]:
+            print("\n  WARNING: Not connected to VTube Studio. Is it running?")
+        if not status["authenticated"]:
+            print("  WARNING: Not authenticated. Run with --auth to initiate auth flow.")
+        return status
+    except Exception as e:
+        print(f"ERROR: Cannot reach bridge at {BRIDGE_URL}: {e}")
+        print("  Is vtube-bridge.py running?")
+        sys.exit(1)
+
+
+def run_auth():
+    """Initiate auth flow — user must click Allow in VTube Studio."""
+    print("Requesting authentication token...")
+    print("  >>> Click 'Allow' in VTube Studio when prompted <<<")
+    result = _post("/auth")
+    print(f"  Result: {json.dumps(result, indent=2)}")
+    return result
+
+
+def test_expressions(delay: float = 2.0):
+    """Cycle through all expressions with a pause between each."""
+    print(f"\nCycling through {len(EXPRESSIONS)} expressions ({delay}s each):\n")
+    for expr in EXPRESSIONS:
+        print(f"  → {expr}...", end=" ", flush=True)
+        t0 = time.monotonic()
+        result = _post("/expression", {"event": expr})
+        dt = (time.monotonic() - t0) * 1000
+        if result.get("ok"):
+            print(f"OK ({dt:.0f}ms)")
+        else:
+            print(f"FAILED: {result.get('error', 'unknown')}")
+        time.sleep(delay)
+
+    # Return to idle
+    _post("/expression", {"event": "idle"})
+    print("\n  Returned to idle.")
+
+
+def test_lipsync(duration: float = 3.0):
+    """Simulate lip sync by sweeping MouthOpen 0→1→0."""
+    import math
+    print(f"\nTesting lip sync (MouthOpen sweep, {duration}s)...\n")
+
+    fps = 20
+    frames = int(duration * fps)
+    for i in range(frames):
+        t = i / frames
+        # Sine wave for smooth open/close
+        value = abs(math.sin(t * math.pi * 4))
+        value = round(value, 3)
+        _post("/parameter", {"name": "MouthOpen", "value": value})
+        print(f"\r  MouthOpen = {value:.3f}", end="", flush=True)
+        time.sleep(1.0 / fps)
+
+    _post("/parameter", {"name": "MouthOpen", "value": 0.0})
+    print("\r  MouthOpen = 0.000 (done)          ")
+
+
+def test_latency(iterations: int = 20):
+    """Measure expression trigger round-trip latency."""
+    print(f"\nMeasuring latency ({iterations} iterations)...\n")
+    times = []
+    for i in range(iterations):
+        expr = "thinking" if i % 2 == 0 else "idle"
+        t0 = time.monotonic()
+        _post("/expression", {"event": expr})
+        dt = (time.monotonic() - t0) * 1000
+        times.append(dt)
+        print(f"  {i+1:2d}. {expr:10s} → {dt:.1f}ms")
+
+    avg = sum(times) / len(times)
+    mn = min(times)
+    mx = max(times)
+    print(f"\n  Avg: {avg:.1f}ms  Min: {mn:.1f}ms  Max: {mx:.1f}ms")
+    if avg < 100:
+        print("  PASS: Average latency under 100ms target")
+    else:
+        print("  WARNING: Average latency exceeds 100ms target")
+
+    # Return to idle
+    _post("/expression", {"event": "idle"})
+
+
+def main():
+    parser = argparse.ArgumentParser(description="VTube Studio Expression Bridge Tester")
+    parser.add_argument("--auth", action="store_true", help="Run auth flow")
+    parser.add_argument("--lipsync", action="store_true", help="Test lip sync parameter sweep")
+    parser.add_argument("--latency", action="store_true", help="Measure round-trip latency")
+    parser.add_argument("--delay", type=float, default=2.0, help="Delay between expressions (default: 2s)")
+    parser.add_argument("--all", action="store_true", help="Run all tests")
+    args = parser.parse_args()
+
+    print("VTube Studio Expression Bridge Tester")
+    print("=" * 42)
+
+    status = check_bridge()
+
+    if args.auth:
+        run_auth()
+        print()
+        status = check_bridge()
+
+    if not status.get("authenticated") and not args.auth:
+        print("\nNot authenticated — skipping expression tests.")
+        print("Run with --auth to authenticate, or start VTube Studio first.")
+        return
+
+    if args.all:
+        test_expressions(args.delay)
+        test_lipsync()
+        test_latency()
+    elif args.lipsync:
+        test_lipsync()
+    elif args.latency:
+        test_latency()
+    else:
+        test_expressions(args.delay)
+
+
+if __name__ == "__main__":
+    main()

homeai-visual/setup.sh

@@ -1,17 +1,16 @@
 #!/usr/bin/env bash
-# homeai-visual/setup.sh — P7: VTube Studio bridge + Live2D expressions
+# homeai-visual/setup.sh — P7: VTube Studio Expression Bridge
 #
-# Components:
+# Sets up:
-#   - vtube_studio.py  — WebSocket client skill for OpenClaw
+#   - Python venv with websockets
-#   - lipsync.py       — amplitude-based lip sync
+#   - vtube-bridge daemon (HTTP ↔ WebSocket bridge)
-#   - auth.py          — VTube Studio token management
+#   - vtube-ctl CLI (symlinked to PATH)
+#   - launchd service
 #
 # Prerequisites:
 #   - P4 (homeai-agent) — OpenClaw running
 #   - P5 (homeai-character) — aria.json with live2d_expressions set
-#   - macOS: VTube Studio installed (Mac App Store)
+#   - VTube Studio installed (Mac App Store) with WebSocket API enabled
-#   - Linux: N/A — VTube Studio is macOS/Windows/iOS only
-#             Linux dev can test the skill code but not the VTube Studio side
 
 set -euo pipefail
 
@@ -19,42 +18,61 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 REPO_DIR="$(cd "${SCRIPT_DIR}/.." && pwd)"
 source "${REPO_DIR}/scripts/common.sh"
 
-log_section "P7: VTube Studio Bridge"
+VENV_DIR="$HOME/homeai-visual-env"
-detect_platform
+PLIST_SRC="${SCRIPT_DIR}/launchd/com.homeai.vtube-bridge.plist"
+PLIST_DST="$HOME/Library/LaunchAgents/com.homeai.vtube-bridge.plist"
+VTUBE_CTL_SRC="$HOME/.openclaw/skills/vtube-studio/scripts/vtube-ctl"
 
-if [[ "$OS_TYPE" == "linux" ]]; then
+log_section "P7: VTube Studio Expression Bridge"
-  log_warn "VTube Studio is not available on Linux."
+
-  log_warn "This sub-project requires macOS (Mac Mini)."
+# ─── Python venv ──────────────────────────────────────────────────────────────
+
+if [[ ! -d "$VENV_DIR" ]]; then
+    log_info "Creating Python venv at $VENV_DIR..."
+    python3 -m venv "$VENV_DIR"
 fi
 
-# ─── TODO: Implementation ──────────────────────────────────────────────────────
+log_info "Installing dependencies..."
-cat <<'EOF'
+"$VENV_DIR/bin/pip" install --upgrade pip -q
+"$VENV_DIR/bin/pip" install websockets -q
+log_ok "Python venv ready ($(${VENV_DIR}/bin/python3 --version))"
 
-  ┌─────────────────────────────────────────────────────────────────┐
+# ─── vtube-ctl symlink ───────────────────────────────────────────────────────
-  │  P7: homeai-visual — NOT YET IMPLEMENTED                       │
-  │                                                                  │
-  │  macOS only (VTube Studio is macOS/iOS/Windows)                 │
-  │                                                                  │
-  │  Implementation steps:                                           │
-  │  1. Install VTube Studio from Mac App Store                     │
-  │  2. Enable WebSocket API in VTube Studio (Settings → port 8001) │
-  │  3. Source/purchase Live2D model                                │
-  │  4. Create expression hotkeys for 8 states                      │
-  │  5. Implement skills/vtube_studio.py (WebSocket client)         │
-  │  6. Implement skills/lipsync.py (amplitude → MouthOpen param)  │
-  │  7. Implement skills/auth.py (token request + persistence)      │
-  │  8. Register vtube_studio skill with OpenClaw                   │
-  │  9. Update aria.json live2d_expressions with hotkey IDs         │
-  │  10. Test all 8 expression states                               │
-  │                                                                  │
-  │  On Linux: implement Python skills, test WebSocket protocol     │
-  │  with a mock server before connecting to real VTube Studio.     │
-  │                                                                  │
-  │  Interface contracts:                                            │
-  │    VTUBE_WS_URL=ws://localhost:8001                             │
-  └─────────────────────────────────────────────────────────────────┘
 
-EOF
+if [[ -f "$VTUBE_CTL_SRC" ]]; then
+    chmod +x "$VTUBE_CTL_SRC"
+    ln -sf "$VTUBE_CTL_SRC" /opt/homebrew/bin/vtube-ctl
+    log_ok "vtube-ctl symlinked to /opt/homebrew/bin/vtube-ctl"
+else
+    log_warn "vtube-ctl not found at $VTUBE_CTL_SRC — skipping symlink"
+fi
 
-log_info "P7 is not yet implemented. See homeai-visual/PLAN.md for details."
+# ─── launchd service ─────────────────────────────────────────────────────────
-exit 0
+
+if [[ -f "$PLIST_SRC" ]]; then
+    # Unload if already loaded
+    launchctl bootout "gui/$(id -u)/com.homeai.vtube-bridge" 2>/dev/null || true
+
+    cp "$PLIST_SRC" "$PLIST_DST"
+    launchctl bootstrap "gui/$(id -u)" "$PLIST_DST"
+    log_ok "launchd service loaded: com.homeai.vtube-bridge"
+else
+    log_warn "Plist not found at $PLIST_SRC — skipping launchd setup"
+fi
+
+# ─── Status ──────────────────────────────────────────────────────────────────
+
+echo ""
+log_info "VTube Bridge setup complete."
+log_info ""
+log_info "Next steps:"
+log_info "  1. Install VTube Studio from Mac App Store"
+log_info "  2. Enable WebSocket API: Settings > WebSocket API > port 8001"
+log_info "  3. Load a Live2D model"
+log_info "  4. Create expression hotkeys (idle, listening, thinking, speaking, happy, sad, surprised, error)"
+log_info "  5. Run: vtube-ctl auth  (click Allow in VTube Studio)"
+log_info "  6. Run: python3 ${SCRIPT_DIR}/scripts/test-expressions.py --all"
+log_info "  7. Update aria.json with real hotkey UUIDs"
+log_info ""
+log_info "Logs: /tmp/homeai-vtube-bridge.log"
+log_info "Bridge: http://localhost:8002/status"

homeai-visual/vtube-bridge.py

@@ -0,0 +1,454 @@
+#!/usr/bin/env python3
+"""
+VTube Studio Expression Bridge — persistent WebSocket ↔ HTTP bridge.
+
+Maintains a long-lived WebSocket connection to VTube Studio and exposes
+a simple HTTP API so other HomeAI components can trigger expressions and
+inject parameters (lip sync) without managing their own WS connections.
+
+HTTP API (port 8002):
+    POST /expression   {"event": "thinking"}          → trigger hotkey
+    POST /parameter    {"name": "MouthOpen", "value": 0.5} → inject param
+    POST /parameters   [{"name": "MouthOpen", "value": 0.5}, ...]
+    POST /auth         {}                              → request new token
+    GET  /status                                       → connection info
+    GET  /expressions                                  → list available expressions
+
+Requires: pip install websockets
+"""
+
+import argparse
+import asyncio
+import json
+import logging
+import signal
+import sys
+import time
+from http import HTTPStatus
+from pathlib import Path
+
+try:
+    import websockets
+    from websockets.exceptions import ConnectionClosed
+except ImportError:
+    print("ERROR: 'websockets' package required. Install with: pip install websockets", file=sys.stderr)
+    sys.exit(1)
+
+# ---------------------------------------------------------------------------
+# Config
+# ---------------------------------------------------------------------------
+
+DEFAULT_VTUBE_WS_URL = "ws://localhost:8001"
+DEFAULT_HTTP_PORT = 8002
+TOKEN_PATH = Path.home() / ".openclaw" / "vtube_token.json"
+DEFAULT_CHARACTER_PATH = (
+    Path.home() / "gitea" / "homeai" / "homeai-dashboard" / "characters" / "aria.json"
+)
+
+logger = logging.getLogger("vtube-bridge")
+
+# ---------------------------------------------------------------------------
+# VTube Studio WebSocket Client
+# ---------------------------------------------------------------------------
+
+class VTubeClient:
+    """Persistent async WebSocket client for VTube Studio API."""
+
+    def __init__(self, ws_url: str, character_path: Path):
+        self.ws_url = ws_url
+        self.character_path = character_path
+        self._ws = None
+        self._token: str | None = None
+        self._authenticated = False
+        self._current_expression: str | None = None
+        self._connected = False
+        self._request_id = 0
+        self._lock = asyncio.Lock()
+        self._load_token()
+        self._load_character()
+
+    # ── Character config ──────────────────────────────────────────────
+
+    def _load_character(self):
+        """Load expression mappings from character JSON."""
+        self.expression_map: dict[str, str] = {}
+        self.ws_triggers: dict = {}
+        try:
+            if self.character_path.exists():
+                cfg = json.loads(self.character_path.read_text())
+                self.expression_map = cfg.get("live2d_expressions", {})
+                self.ws_triggers = cfg.get("vtube_ws_triggers", {})
+                logger.info("Loaded %d expressions from %s", len(self.expression_map), self.character_path.name)
+            else:
+                logger.warning("Character file not found: %s", self.character_path)
+        except Exception as e:
+            logger.error("Failed to load character config: %s", e)
+
+    def reload_character(self):
+        """Hot-reload character config without restarting."""
+        self._load_character()
+        return {"expressions": self.expression_map, "triggers": self.ws_triggers}
+
+    # ── Token persistence ─────────────────────────────────────────────
+
+    def _load_token(self):
+        try:
+            if TOKEN_PATH.exists():
+                data = json.loads(TOKEN_PATH.read_text())
+                self._token = data.get("token")
+                logger.info("Loaded auth token from %s", TOKEN_PATH)
+        except Exception as e:
+            logger.warning("Could not load token: %s", e)
+
+    def _save_token(self, token: str):
+        TOKEN_PATH.parent.mkdir(parents=True, exist_ok=True)
+        TOKEN_PATH.write_text(json.dumps({"token": token}, indent=2))
+        self._token = token
+        logger.info("Saved auth token to %s", TOKEN_PATH)
+
+    # ── WebSocket comms ───────────────────────────────────────────────
+
+    def _next_id(self) -> str:
+        self._request_id += 1
+        return f"homeai-{self._request_id}"
+
+    async def _send(self, message_type: str, data: dict | None = None) -> dict:
+        """Send a VTube Studio API message and return the response."""
+        payload = {
+            "apiName": "VTubeStudioPublicAPI",
+            "apiVersion": "1.0",
+            "requestID": self._next_id(),
+            "messageType": message_type,
+            "data": data or {},
+        }
+        await self._ws.send(json.dumps(payload))
+        resp = json.loads(await asyncio.wait_for(self._ws.recv(), timeout=10))
+        return resp
+
+    # ── Connection lifecycle ──────────────────────────────────────────
+
+    async def connect(self):
+        """Connect and authenticate to VTube Studio."""
+        try:
+            self._ws = await websockets.connect(self.ws_url, ping_interval=20, ping_timeout=10)
+            self._connected = True
+            logger.info("Connected to VTube Studio at %s", self.ws_url)
+
+            if self._token:
+                await self._authenticate()
+            else:
+                logger.warning("No auth token — call POST /auth to initiate authentication")
+        except Exception as e:
+            self._connected = False
+            self._authenticated = False
+            logger.error("Connection failed: %s", e)
+            raise
+
+    async def _authenticate(self):
+        """Authenticate with an existing token."""
+        resp = await self._send("AuthenticationRequest", {
+            "pluginName": "HomeAI",
+            "pluginDeveloper": "HomeAI",
+            "authenticationToken": self._token,
+        })
+        self._authenticated = resp.get("data", {}).get("authenticated", False)
+        if self._authenticated:
+            logger.info("Authenticated successfully")
+        else:
+            logger.warning("Token rejected — request a new one via POST /auth")
+            self._authenticated = False
+
+    async def request_new_token(self) -> dict:
+        """Request a new auth token. User must click Allow in VTube Studio."""
+        if not self._connected:
+            return {"error": "Not connected to VTube Studio"}
+
+        resp = await self._send("AuthenticationTokenRequest", {
+            "pluginName": "HomeAI",
+            "pluginDeveloper": "HomeAI",
+            "pluginIcon": None,
+        })
+        token = resp.get("data", {}).get("authenticationToken")
+        if token:
+            self._save_token(token)
+            await self._authenticate()
+            return {"authenticated": self._authenticated, "token_saved": True}
+        return {"error": "No token received", "response": resp}
+
+    async def disconnect(self):
+        if self._ws:
+            await self._ws.close()
+        self._connected = False
+        self._authenticated = False
+
+    async def ensure_connected(self):
+        """Reconnect if the connection dropped."""
+        if not self._connected or self._ws is None or self._ws.closed:
+            logger.info("Reconnecting...")
+            await self.connect()
+
+    # ── Expression & parameter API ────────────────────────────────────
+
+    async def trigger_expression(self, event: str) -> dict:
+        """Trigger a named expression from the character config."""
+        async with self._lock:
+            await self.ensure_connected()
+            if not self._authenticated:
+                return {"error": "Not authenticated"}
+
+            hotkey_id = self.expression_map.get(event)
+            if not hotkey_id:
+                return {"error": f"Unknown expression: {event}", "available": list(self.expression_map.keys())}
+
+            resp = await self._send("HotkeyTriggerRequest", {"hotkeyID": hotkey_id})
+            self._current_expression = event
+            return {"ok": True, "expression": event, "hotkey_id": hotkey_id}
+
+    async def set_parameter(self, name: str, value: float, weight: float = 1.0) -> dict:
+        """Inject a single VTube Studio parameter value."""
+        async with self._lock:
+            await self.ensure_connected()
+            if not self._authenticated:
+                return {"error": "Not authenticated"}
+
+            resp = await self._send("InjectParameterDataRequest", {
+                "parameterValues": [{"id": name, "value": value, "weight": weight}],
+            })
+            return {"ok": True, "name": name, "value": value}
+
+    async def set_parameters(self, params: list[dict]) -> dict:
+        """Inject multiple VTube Studio parameters at once."""
+        async with self._lock:
+            await self.ensure_connected()
+            if not self._authenticated:
+                return {"error": "Not authenticated"}
+
+            param_values = [
+                {"id": p["name"], "value": p["value"], "weight": p.get("weight", 1.0)}
+                for p in params
+            ]
+            resp = await self._send("InjectParameterDataRequest", {
+                "parameterValues": param_values,
+            })
+            return {"ok": True, "count": len(param_values)}
+
+    async def list_hotkeys(self) -> dict:
+        """List all hotkeys available in the current model."""
+        async with self._lock:
+            await self.ensure_connected()
+            if not self._authenticated:
+                return {"error": "Not authenticated"}
+
+            resp = await self._send("HotkeysInCurrentModelRequest", {})
+            return resp.get("data", {})
+
+    async def list_parameters(self) -> dict:
+        """List all input parameters for the current model."""
+        async with self._lock:
+            await self.ensure_connected()
+            if not self._authenticated:
+                return {"error": "Not authenticated"}
+
+            resp = await self._send("InputParameterListRequest", {})
+            return resp.get("data", {})
+
+    def status(self) -> dict:
+        return {
+            "connected": self._connected,
+            "authenticated": self._authenticated,
+            "ws_url": self.ws_url,
+            "current_expression": self._current_expression,
+            "expression_count": len(self.expression_map),
+            "expressions": list(self.expression_map.keys()),
+        }
+
+
+# ---------------------------------------------------------------------------
+# HTTP Server (asyncio-based, no external deps)
+# ---------------------------------------------------------------------------
+
+class BridgeHTTPHandler:
+    """Simple async HTTP request handler for the bridge API."""
+
+    def __init__(self, client: VTubeClient):
+        self.client = client
+
+    async def handle(self, reader: asyncio.StreamReader, writer: asyncio.StreamWriter):
+        try:
+            request_line = await asyncio.wait_for(reader.readline(), timeout=5)
+            if not request_line:
+                writer.close()
+                return
+
+            method, path, _ = request_line.decode().strip().split(" ", 2)
+            path = path.split("?")[0]  # strip query params
+
+            # Read headers
+            content_length = 0
+            while True:
+                line = await reader.readline()
+                if line == b"\r\n" or not line:
+                    break
+                if line.lower().startswith(b"content-length:"):
+                    content_length = int(line.split(b":")[1].strip())
+
+            # Read body
+            body = None
+            if content_length > 0:
+                body = await reader.read(content_length)
+
+            # Route
+            try:
+                result = await self._route(method, path, body)
+                await self._respond(writer, 200, result)
+            except Exception as e:
+                logger.error("Handler error: %s", e, exc_info=True)
+                await self._respond(writer, 500, {"error": str(e)})
+
+        except asyncio.TimeoutError:
+            writer.close()
+        except Exception as e:
+            logger.error("Connection error: %s", e)
+            try:
+                writer.close()
+            except Exception:
+                pass
+
+    async def _route(self, method: str, path: str, body: bytes | None) -> dict:
+        data = {}
+        if body:
+            try:
+                data = json.loads(body)
+            except json.JSONDecodeError:
+                return {"error": "Invalid JSON"}
+
+        if method == "GET" and path == "/status":
+            return self.client.status()
+
+        if method == "GET" and path == "/expressions":
+            return {
+                "expressions": self.client.expression_map,
+                "triggers": self.client.ws_triggers,
+            }
+
+        if method == "GET" and path == "/hotkeys":
+            return await self.client.list_hotkeys()
+
+        if method == "GET" and path == "/parameters":
+            return await self.client.list_parameters()
+
+        if method == "POST" and path == "/expression":
+            event = data.get("event")
+            if not event:
+                return {"error": "Missing 'event' field"}
+            return await self.client.trigger_expression(event)
+
+        if method == "POST" and path == "/parameter":
+            name = data.get("name")
+            value = data.get("value")
+            if name is None or value is None:
+                return {"error": "Missing 'name' or 'value' field"}
+            return await self.client.set_parameter(name, float(value), float(data.get("weight", 1.0)))
+
+        if method == "POST" and path == "/parameters":
+            if not isinstance(data, list):
+                return {"error": "Expected JSON array of {name, value} objects"}
+            return await self.client.set_parameters(data)
+
+        if method == "POST" and path == "/auth":
+            return await self.client.request_new_token()
+
+        if method == "POST" and path == "/reload":
+            return self.client.reload_character()
+
+        return {"error": f"Unknown route: {method} {path}"}
+
+    async def _respond(self, writer: asyncio.StreamWriter, status: int, data: dict):
+        body = json.dumps(data, indent=2).encode()
+        status_text = HTTPStatus(status).phrase
+        header = (
+            f"HTTP/1.1 {status} {status_text}\r\n"
+            f"Content-Type: application/json\r\n"
+            f"Content-Length: {len(body)}\r\n"
+            f"Access-Control-Allow-Origin: *\r\n"
+            f"Access-Control-Allow-Methods: GET, POST, OPTIONS\r\n"
+            f"Access-Control-Allow-Headers: Content-Type\r\n"
+            f"\r\n"
+        )
+        writer.write(header.encode() + body)
+        await writer.drain()
+        writer.close()
+
+
+# ---------------------------------------------------------------------------
+# Auto-reconnect loop
+# ---------------------------------------------------------------------------
+
+async def reconnect_loop(client: VTubeClient, interval: float = 5.0):
+    """Background task that keeps the VTube Studio connection alive."""
+    while True:
+        try:
+            if not client._connected or client._ws is None or client._ws.closed:
+                logger.info("Connection lost — attempting reconnect...")
+                await client.connect()
+        except Exception as e:
+            logger.debug("Reconnect failed: %s (retrying in %.0fs)", e, interval)
+        await asyncio.sleep(interval)
+
+
+# ---------------------------------------------------------------------------
+# Main
+# ---------------------------------------------------------------------------
+
+async def main(args):
+    logging.basicConfig(
+        level=logging.DEBUG if args.verbose else logging.INFO,
+        format="%(asctime)s [%(name)s] %(levelname)s: %(message)s",
+        datefmt="%H:%M:%S",
+    )
+
+    character_path = Path(args.character)
+    client = VTubeClient(args.vtube_url, character_path)
+
+    # Try initial connection (don't fail if VTube Studio isn't running yet)
+    try:
+        await client.connect()
+    except Exception as e:
+        logger.warning("Initial connection failed: %s (will keep retrying)", e)
+
+    # Start reconnect loop
+    reconnect_task = asyncio.create_task(reconnect_loop(client, interval=5.0))
+
+    # Start HTTP server
+    handler = BridgeHTTPHandler(client)
+    server = await asyncio.start_server(handler.handle, "0.0.0.0", args.port)
+    logger.info("HTTP API listening on http://0.0.0.0:%d", args.port)
+    logger.info("Endpoints: /status /expression /parameter /parameters /auth /reload /hotkeys")
+
+    # Graceful shutdown
+    stop = asyncio.Event()
+
+    def _signal_handler():
+        logger.info("Shutting down...")
+        stop.set()
+
+    loop = asyncio.get_event_loop()
+    for sig in (signal.SIGINT, signal.SIGTERM):
+        loop.add_signal_handler(sig, _signal_handler)
+
+    async with server:
+        await stop.wait()
+
+    reconnect_task.cancel()
+    await client.disconnect()
+    logger.info("Goodbye.")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="VTube Studio Expression Bridge")
+    parser.add_argument("--port", type=int, default=DEFAULT_HTTP_PORT, help="HTTP API port (default: 8002)")
+    parser.add_argument("--vtube-url", default=DEFAULT_VTUBE_WS_URL, help="VTube Studio WebSocket URL")
+    parser.add_argument("--character", default=str(DEFAULT_CHARACTER_PATH), help="Path to character JSON")
+    parser.add_argument("--verbose", "-v", action="store_true", help="Debug logging")
+    args = parser.parse_args()
+    asyncio.run(main(args))

homeai-voice/scripts/launchd/com.homeai.wyoming-tts.plist

@@ -18,6 +18,12 @@
     <string>1.0</string>
   </array>
 
+  <key>EnvironmentVariables</key>
+  <dict>
+    <key>ELEVENLABS_API_KEY</key>
+    <string>sk_ec10e261c6190307a37aa161a9583504dcf25a0cabe5dbd5</string>
+  </dict>
+
   <key>RunAtLoad</key>
   <true/>
 

homeai-voice/tts/wyoming_kokoro_server.py

@@ -7,8 +7,10 @@ Usage:
 
 import argparse
 import asyncio
+import json
 import logging
 import os
+import urllib.request
 
 import numpy as np
 
@@ -20,10 +22,76 @@ from wyoming.tts import Synthesize
 
 _LOGGER = logging.getLogger(__name__)
 
+ACTIVE_TTS_VOICE_PATH = os.path.expanduser("~/homeai-data/active-tts-voice.json")
 SAMPLE_RATE = 24000
 SAMPLE_WIDTH = 2  # int16
 CHANNELS = 1
 CHUNK_SECONDS = 1  # stream in 1-second chunks
+VTUBE_BRIDGE_URL = "http://localhost:8002"
+LIPSYNC_ENABLED = True
+LIPSYNC_FRAME_SAMPLES = 1200  # 50ms frames at 24kHz → 20 updates/sec
+LIPSYNC_SCALE = 10.0  # amplitude multiplier (tuned for Kokoro output levels)
+
+
+def _send_lipsync(value: float):
+    """Fire-and-forget POST to vtube-bridge with mouth open value."""
+    try:
+        body = json.dumps({"name": "MouthOpen", "value": value}).encode()
+        req = urllib.request.Request(
+            f"{VTUBE_BRIDGE_URL}/parameter",
+            data=body,
+            headers={"Content-Type": "application/json"},
+            method="POST",
+        )
+        urllib.request.urlopen(req, timeout=0.5)
+    except Exception:
+        pass  # bridge may not be running
+
+
+def _compute_lipsync_frames(samples_int16: np.ndarray) -> list[float]:
+    """Compute per-frame RMS amplitude scaled to 0–1 for lip sync."""
+    frames = []
+    for i in range(0, len(samples_int16), LIPSYNC_FRAME_SAMPLES):
+        frame = samples_int16[i : i + LIPSYNC_FRAME_SAMPLES].astype(np.float32)
+        rms = np.sqrt(np.mean(frame ** 2)) / 32768.0
+        mouth = min(rms * LIPSYNC_SCALE, 1.0)
+        frames.append(round(mouth, 3))
+    return frames
+
+
+def _get_active_tts_config() -> dict | None:
+    """Read the active TTS config set by the OpenClaw bridge."""
+    try:
+        with open(ACTIVE_TTS_VOICE_PATH) as f:
+            return json.load(f)
+    except Exception:
+        return None
+
+
+def _synthesize_elevenlabs(text: str, voice_id: str, model: str = "eleven_multilingual_v2") -> bytes:
+    """Call ElevenLabs TTS API and return raw PCM audio bytes (24kHz 16-bit mono)."""
+    api_key = os.environ.get("ELEVENLABS_API_KEY", "")
+    if not api_key:
+        raise RuntimeError("ELEVENLABS_API_KEY not set")
+
+    url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}?output_format=pcm_24000"
+    payload = json.dumps({
+        "text": text,
+        "model_id": model,
+        "voice_settings": {"stability": 0.5, "similarity_boost": 0.75},
+    }).encode()
+
+    req = urllib.request.Request(
+        url,
+        data=payload,
+        headers={
+            "Content-Type": "application/json",
+            "xi-api-key": api_key,
+        },
+        method="POST",
+    )
+    with urllib.request.urlopen(req, timeout=30) as resp:
+        return resp.read()
 
 
 def _load_kokoro():
@@ -76,26 +144,53 @@ class KokoroEventHandler(AsyncEventHandler):
             synthesize = Synthesize.from_event(event)
             text = synthesize.text
             voice = self._default_voice
+            use_elevenlabs = False
 
-            if synthesize.voice and synthesize.voice.name:
+            # Bridge state file takes priority (set per-request by OpenClaw bridge)
+            tts_config = _get_active_tts_config()
+            if tts_config and tts_config.get("engine") == "elevenlabs":
+                use_elevenlabs = True
+                voice = tts_config.get("elevenlabs_voice_id", "")
+                _LOGGER.debug("Synthesizing %r with ElevenLabs voice=%s", text, voice)
+            elif tts_config and tts_config.get("kokoro_voice"):
+                voice = tts_config["kokoro_voice"]
+            elif synthesize.voice and synthesize.voice.name:
                 voice = synthesize.voice.name
 
-            _LOGGER.debug("Synthesizing %r with voice=%s speed=%.1f", text, voice, self._speed)
-
             try:
                 loop = asyncio.get_event_loop()
-                samples, sample_rate = await loop.run_in_executor(
-                    None, lambda: self._tts.create(text, voice=voice, speed=self._speed)
-                )
 
-                samples_int16 = (np.clip(samples, -1.0, 1.0) * 32767).astype(np.int16)
+                if use_elevenlabs and voice:
-                audio_bytes = samples_int16.tobytes()
+                    # ElevenLabs returns PCM 24kHz 16-bit mono
+                    model = tts_config.get("elevenlabs_model", "eleven_multilingual_v2")
+                    _LOGGER.info("Using ElevenLabs TTS (model=%s, voice=%s)", model, voice)
+                    pcm_bytes = await loop.run_in_executor(
+                        None, lambda: _synthesize_elevenlabs(text, voice, model)
+                    )
+                    samples_int16 = np.frombuffer(pcm_bytes, dtype=np.int16)
+                    audio_bytes = pcm_bytes
+                else:
+                    _LOGGER.debug("Synthesizing %r with Kokoro voice=%s speed=%.1f", text, voice, self._speed)
+                    samples, sample_rate = await loop.run_in_executor(
+                        None, lambda: self._tts.create(text, voice=voice, speed=self._speed)
+                    )
+                    samples_int16 = (np.clip(samples, -1.0, 1.0) * 32767).astype(np.int16)
+                    audio_bytes = samples_int16.tobytes()
+
+                # Pre-compute lip sync frames for the entire utterance
+                lipsync_frames = []
+                if LIPSYNC_ENABLED:
+                    lipsync_frames = _compute_lipsync_frames(samples_int16)
 
                 await self.write_event(
                     AudioStart(rate=SAMPLE_RATE, width=SAMPLE_WIDTH, channels=CHANNELS).event()
                 )
 
                 chunk_size = SAMPLE_RATE * SAMPLE_WIDTH * CHANNELS * CHUNK_SECONDS
+                lipsync_idx = 0
+                samples_per_chunk = SAMPLE_RATE * CHUNK_SECONDS
+                frames_per_chunk = samples_per_chunk // LIPSYNC_FRAME_SAMPLES
+
                 for i in range(0, len(audio_bytes), chunk_size):
                     await self.write_event(
                         AudioChunk(
@@ -106,8 +201,22 @@ class KokoroEventHandler(AsyncEventHandler):
                         ).event()
                     )
 
+                    # Send lip sync frames for this audio chunk
+                    if LIPSYNC_ENABLED and lipsync_frames:
+                        chunk_frames = lipsync_frames[lipsync_idx : lipsync_idx + frames_per_chunk]
+                        for mouth_val in chunk_frames:
+                            await asyncio.get_event_loop().run_in_executor(
+                                None, _send_lipsync, mouth_val
+                            )
+                        lipsync_idx += frames_per_chunk
+
+                # Close mouth after speech
+                if LIPSYNC_ENABLED:
+                    await asyncio.get_event_loop().run_in_executor(None, _send_lipsync, 0.0)
+
                 await self.write_event(AudioStop().event())
-                _LOGGER.info("Synthesized %.1fs of audio", len(samples) / sample_rate)
+                duration = len(samples_int16) / SAMPLE_RATE
+                _LOGGER.info("Synthesized %.1fs of audio (%d lipsync frames)", duration, len(lipsync_frames))
 
             except Exception:
                 _LOGGER.exception("Synthesis error")