https://github.com/outspeed-ai/voice-devtools
Developer tools to debug and build realtime voice agents. Supports multiple models.
https://github.com/outspeed-ai/voice-devtools
javascript multimodal nodejs open-source voice-ai webrtc
Last synced: about 1 year ago
JSON representation
Developer tools to debug and build realtime voice agents. Supports multiple models.
- Host: GitHub
- URL: https://github.com/outspeed-ai/voice-devtools
- Owner: outspeed-ai
- License: mit
- Created: 2025-02-21T05:06:23.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-04-11T11:38:04.000Z (about 1 year ago)
- Last Synced: 2025-04-12T04:06:25.459Z (about 1 year ago)
- Topics: javascript, multimodal, nodejs, open-source, voice-ai, webrtc
- Language: TypeScript
- Homepage: https://demo.outspeed.com
- Size: 2.15 MB
- Stars: 26
- Watchers: 2
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Voice DevTools
This UI provides a debug console for real-time AI voice interactions. It works with multiple realtime models ([View supported models](https://docs.outspeed.com/models)). Features include:
1. Cost Tracking: Know how much you've spent per voice interaction
2. Model Support: Supports open-source (MiniCPM-o) and closed-source S2S models (adding more soon!)
3. Metrics Support: Track errors, response timelines, and performance
4. Voice and Chat UI
Inspired by [openai-realtime-console](https://github.com/openai/openai-realtime-console) and [openai-realtime-agents](https://github.com/openai/openai-realtime-agents).
## Quick Start
1. Get your API keys:
- [Outspeed API key](https://dashboard.outspeed.com)
- [OpenAI API key](https://platform.openai.com/settings/api-keys)
2. Set up environment:
```bash
cp .env.example .env
# Add your API keys to .env:
# OPENAI_API_KEY=""
# OUTSPEED_API_KEY=""
```
3. Install and run:
```bash
npm install
npm run dev
```
Visit [http://localhost:3000](http://localhost:3000) to access the console.
### Usage
To modify agent prompt and tools, modify `agent-config.js`.
To modify the model parameters like (voice, version, etc.), edit `model-config.js`
### Agents
There are two voice agent examples already present in `agent-config.js`:
1. Dental Agent: Answers callers' questions about working hours of a dental clinic
2. Message Agent: Takes callers' messages for a person
You can modify the export in the file to activate an agent:
```js
// agent-config.js
export { dental_agent as agent }
// for message_agent
export { message_agent as agent }
```
Play around with the prompts, or add your own voice agent to give it a spin.
### Supported Models
- [MiniCPM-o](https://github.com/OpenBMB/MiniCPM-o) (hosted by Outspeed)
- OpenAI Realtime models
- Moshi (Coming Soon)
- Gemini Multimodal Live (Coming Soon)
## License
MIT