https://github.com/composiohq/voice-website-agent
https://github.com/composiohq/voice-website-agent
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/composiohq/voice-website-agent
- Owner: ComposioHQ
- Created: 2025-08-11T14:28:07.000Z (about 2 months ago)
- Default Branch: main
- Last Pushed: 2025-08-12T15:26:53.000Z (about 2 months ago)
- Last Synced: 2025-08-12T17:21:45.412Z (about 2 months ago)
- Language: TypeScript
- Size: 54.7 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## Voice Website Agent
Speak your intent. The agent transcribes your voice, plans with GPT‑5, executes tools to update a live HTML preview, and replies by voice.
### Features
- Modular pipeline: STT → Agent (with tools) → TTS
- Progressive UX: chat updates after each stage
- Composio tool `WRITE_FULL_HTML_PREVIEW` to write `public/preview.html`
- Composio Notion tools: `NOTION_FETCH_DATA`, `NOTION_FETCH_BLOCK_CONTENTS`
- Next.js App Router (Node runtime)
### Quickstart
1) Install
```bash
npm install
```
2) Environment
Create `.env.local`:
```bash
OPENAI_API_KEY=your_openai_api_key
COMPOSIO_API_KEY=your_composio_api_key
```
3) Run
```bash
npm run dev
```
Open http://localhost:3000 and click the mic.
### API routes
- POST `/api/stt`
- Request: multipart form with `audio` (e.g., `audio/webm`)
- Response: `{ transcript: string }`
- POST `/api/agent`
- Request: JSON `{ text: string }`
- Response: `{ text: string }`
- POST `/api/tts`
- Request: JSON `{ text: string }`
- Response: `{ audioBase64: string, mimeType: string }`
- POST `/api/voice-chat` (orchestrator)
- Request: multipart form with `audio`
- Response: `{ transcript, text, audioBase64, mimeType }`
### Architecture
- `app/api/_lib/agentState.ts`: minimal global message state + OpenAI message adapter
- `app/api/_lib/openaiClient.ts`: OpenAI client factory
- `app/api/_lib/composioTools.ts`: Composio setup and `ensureToolsRegistered()`
- `app/api/_lib/stt.ts`: Whisper transcription
- `app/api/_lib/agent.ts`: GPT‑5 reasoning, tool calls, summary + voice‑friendly rewrite
- `app/api/_lib/tts.ts`: Text‑to‑speech with length‑safe condensation
Notes
- Global state is in‑memory; use persistent storage for multi‑user production.
- Tool list is intentionally small to keep behavior focused and safe.
### cURL examples
STT
```bash
curl -X POST http://localhost:3000/api/stt \
-F "audio=@input.webm;type=audio/webm"
```
Agent
```bash
curl -X POST http://localhost:3000/api/agent \
-H "Content-Type: application/json" \
-d '{"text":"Create a landing page with a hero and CTA"}'
```
TTS
```bash
curl -X POST http://localhost:3000/api/tts \
-H "Content-Type: application/json" \
-d '{"text":"Your page is ready."}'
```
### Composio
- GitHub: [Composio](https://github.com/composiohq/composio)
- Website: [composio.dev](https://composio.dev)
### Composio + Notion setup
1) Create API key
- Go to the Composio platform dashboard and create an API key.
- Add it to `.env.local` as `COMPOSIO_API_KEY` (see Environment section above).
2) Connect Notion
- In the Composio platform, Add Notion to your project.
- Complete the Notion OAuth flow. Share the pages/databases you want the agent to read with the Composio integration so it has access.
- Ensure this Notion connection is available as the default connection (aka default user/connection). The agent will use the first connected account by default;
3) Tools used
- `NOTION_FETCH_DATA`: fetches metadata/data for pages or databases.
- `NOTION_FETCH_BLOCK_CONTENTS`: fetches the block contents of a page.