Add voice-controlled home agent with skills and guardrails

.env.example

@@ -0,0 +1,17 @@
+# Gemini API (required for voice streaming)
+GOOGLE_GENAI_USE_VERTEXAI=FALSE
+GOOGLE_API_KEY=your_gemini_api_key
+
+# Voice model must support Gemini Live API
+HOME_AGENT_MODEL=gemini-2.0-flash-live-001
+
+# Home Assistant
+HOME_ASSISTANT_URL=http://homeassistant.local:8123
+HOME_ASSISTANT_TOKEN=your_long_lived_access_token
+
+# Voice server
+VOICE_SERVER_HOST=0.0.0.0
+VOICE_SERVER_PORT=8000
+
+# Optional: local Ollama for text-only dev with doc_agent
+OLLAMA_API_BASE=http://localhost:11434

.gitignore

@@ -58,8 +58,12 @@ ENV/
 env.bak/
 venv.bak/
 
-# macOS
+# macOS / Xcode
 .DS_Store
+*.xcuserstate
+xcuserdata/
+DerivedData/
+build/
 
 # VSCode
 .vscode/

Makefile

@@ -28,6 +28,14 @@ test: ## Run all tests
 		--cov-report=term-missing \
 		--junitxml=junit.xml
 
+.PHONY: voice
+voice: ## Run the iPhone-friendly voice server
+	@PYTHONPATH=src/agents:src python3 -m server.voice_server
+
+.PHONY: home_web
+home_web: ## Run ADK web UI for the home agent
+	@uv run adk web --reload src/agents/home_agent
+
 .PHONY: web
 web: ## Run the ADK web demo server
 	@uv run adk web --reload src/agents/

README.md

@@ -24,6 +24,11 @@ make prefect-server
 - **`doc_agent`**: Technical writing assistant with Google Docs integration
   - Uses local Ollama models (`gpt-oss:20b`) via LiteLLM
   - Includes Google Docs tools for creating and formatting documents
+- **`home_agent`**: Voice-controlled home assistant for iPhone and browsers
+  - Uses Gemini Live API for real-time voice conversations
+  - Pluggable skills (Home Assistant, routines) via `config/skills.yaml`
+  - Safety guardrails via `config/guardrails.yaml`
+  - Native iOS app (`ios/`) and web PWA served by the voice WebSocket server
 
 ### Workflows
 - **`pipeline.py`**: Prefect workflow for orchestrating document creation and content generation
@@ -45,6 +50,8 @@ make prefect-server
 
 ```bash
 make web              # Run ADK web interface
+make home_web         # Run ADK web UI for home agent only
+make voice            # Run iPhone-friendly voice server
 make api_server       # Run ADK FastAPI server
 make prefect-server   # Start Prefect server and serve flows
 make prefect-flows    # Serve flows (server must be running)
@@ -57,18 +64,52 @@ make check            # Lint and type check
 ```
 src/
 ├── agents/
-│   └── doc_agent/           # Technical writing assistant
-│       ├── agent.py         # Agent definition
+│   ├── doc_agent/           # Technical writing assistant
+│   │   ├── agent.py
+│   │   └── tools/
+│   └── home_agent/          # Voice home assistant
+│       ├── agent.py
+│       ├── config/          # skills.yaml, guardrails.yaml
+│       ├── guardrails/
+│       ├── skills/
 │       └── tools/
-│           └── google_docs_tool.py  # Google Docs integration
+├── server/
+│   ├── voice_server.py      # WebSocket voice server + PWA
+│   └── static/              # Browser web client
 └── workflows/
-    ├── pipeline.py          # Main Prefect workflow
-    ├── serve.py             # Flow serving entry point
-    └── discover.py          # Flow discovery utility
+ios/
+└── HomeVoiceAgent.xcodeproj # Native iPhone app (SwiftUI)
 ```
 
 ## Configuration
 
 - **Ollama**: Set `OLLAMA_API_BASE` in `.env` or environment
 - **Google Docs**: Place `credentials.json` in project root (token.json auto-generated)
+- **Home Voice Agent**: Copy `.env.example` to `.env` and set:
+  - `GOOGLE_API_KEY` — Gemini API key (required for voice)
+  - `HOME_ASSISTANT_URL` and `HOME_ASSISTANT_TOKEN` — Home Assistant access
+  - Customize skills in `src/agents/home_agent/config/skills.yaml`
+  - Customize guardrails in `src/agents/home_agent/config/guardrails.yaml`
 - **Prefect UI**: Available at `http://127.0.0.1:4200` (or port specified by `PREFECT_PORT`)
+
+## Home Voice Agent (iPhone)
+
+### Native iOS app (recommended)
+
+1. Copy `.env.example` to `.env` and add your Gemini + Home Assistant credentials
+2. Start the voice server on your Mac: `make voice`
+3. Open `ios/HomeVoiceAgent.xcodeproj` in Xcode
+4. Set your development team, build to your iPhone
+5. In app Settings, enter your server host (e.g. `192.168.1.42:8000`)
+6. Tap **Start Voice** and talk to your home
+
+See `ios/README.md` for full setup and remote access options.
+
+### Web client (alternative)
+
+1. Start the voice server: `make voice`
+2. On your iPhone (same Wi‑Fi), open `http://<your-computer-ip>:8000`
+3. Tap **Start Voice**, allow the microphone, and talk to your home
+
+For HTTPS/WSS from your phone outside the LAN, put the server behind a reverse proxy
+or tunnel (e.g. Cloudflare Tunnel, Tailscale) and use `wss://`.