{"id":28750503,"url":"https://github.com/allwin107/ai-voice-agent","last_synced_at":"2025-06-16T21:39:23.350Z","repository":{"id":298258085,"uuid":"999369127","full_name":"allwin107/AI-Voice-Agent","owner":"allwin107","description":"🎙️ AI Voice Agent is a multi-language voice assistant that transcribes speech, generates intelligent responses using LLaMA 3 via Groq, and replies with realistic synthesized speech using ElevenLabs.","archived":false,"fork":false,"pushed_at":"2025-06-10T07:17:17.000Z","size":0,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-06-10T07:33:47.693Z","etag":null,"topics":["agent","ai","ai-agent","elevenlabs","gorq","llm","python","speech-to-text","text-to-speech","voice-agent"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/allwin107.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-06-10T06:38:21.000Z","updated_at":"2025-06-10T07:22:32.000Z","dependencies_parsed_at":"2025-06-10T07:33:49.210Z","dependency_job_id":"d44fa24d-daba-4f7b-9b21-1fd8f6d3eb57","html_url":"https://github.com/allwin107/AI-Voice-Agent","commit_stats":null,"previous_names":["allwin107/ai-voice-agent"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/allwin107/AI-Voice-Agent","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/allwin107%2FAI-Voice-Agent","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/allwin107%2FAI-Voice-Agent/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/allwin107%2FAI-Voice-Agent/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/allwin107%2FAI-Voice-Agent/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/allwin107","download_url":"https://codeload.github.com/allwin107/AI-Voice-Agent/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/allwin107%2FAI-Voice-Agent/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":260246454,"owners_count":22980395,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent","ai","ai-agent","elevenlabs","gorq","llm","python","speech-to-text","text-to-speech","voice-agent"],"created_at":"2025-06-16T21:39:21.753Z","updated_at":"2025-06-16T21:39:23.337Z","avatar_url":"https://github.com/allwin107.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# AI Voice Agent - proPAL AI - Assignment\n\nA real-time voice interaction system built with LiveKit that combines Speech-to-Text, Large Language Model, and Text-to-Speech capabilities to create an interactive voice agent.\n\n## Features\n\n- Speech-to-Text (STT) using OpenAI's Whisper\n- Large Language Model (LLM) integration with Groq (Llama3-70B model)\n- Text-to-Speech (TTS) using ElevenLabs\n- Real-time streaming support via LiveKit\n- Comprehensive metrics tracking and logging to Excel\n- Multi-language support\n\n## Quick Start\n\n### 1. Prerequisites\n\n- Python 3.8 or higher\n- Virtual environment (recommended)\n\n### 2. Installation\n\n```bash\n# Clone or download the project files\ngit clone https://github.com/allwin107/AI-Voice-Agent.git\ncd ai-voice-agent\n\n# Create virtual environment\npython -m venv venv\nsource venv/bin/activate  # On Windows: venv\\Scripts\\activate\n```\n\n### Requirements\n\nInstall the required dependencies:\n\n```bash\npip install -r requirements.txt\n```\n\n## Configuration\n\nCreate a `.env` file with your API keys:\n\n```bash\nGROQ_API_KEY=your_groq_api_key\nELEVENLABS_API_KEY=your_elevenlabs_api_key\n\n# LiveKit Configuration\nLIVEKIT_URL=wss://your-project.livekit.cloud\nLIVEKIT_API_KEY=your_api_key\nLIVEKIT_API_SECRET=your_api_secret\n```\n### Getting API Keys\n\n1. **Gorq (free LLM)** [console.groq.com](https://console.groq.com) - Fast LLM inference\n2. **ElevenLabs**: [elevenlabs.io](https://elevenlabs.io/) - Text-to-Speech\n3. **Livekit**: [LiveKit Cloud](https://cloud.livekit.io/) - Real-time communication\n\n\n## Project Structure\n\n- `app/pipeline/` - Core pipeline components\n  - `stt.py` - Speech-to-Text using Whisper\n  - `llm.py` - Language model integration using Groq\n  - `tts.py` - Text-to-Speech using ElevenLabs\n  - `voice_agent.py` - Main voice agent pipeline\n  - `livekit_backend.py` - Livekit Integration\n- `app/test/` - Testing Scripts\n  - `test_stt.py` - Tests the transcription functionality of the STT (Speech-to-Text) module.\n  - `test_llm.py` - This script is used to test the LLM response generation functionality.\n  - `test_tts.py` - Test the text-to-speech functionality of the application.\n  - `test_agent.py` - Test script for the voice agent pipeline\n  - `test_audio` - Test .wav audio file\n- `app/config.py` - Configuration settings for the application\n- `.env` - Environment Variables\n- `README.md`\n- `requirements.txt`\n\n## Usage\n\n### Running Tests\n\nTest individual components:\n\n```bash\npython app/pipeline/test_stt.py\npython app/pipeline/test_llm.py\npython app/pipeline/test_tts.py\npython app/pipeline/test_agent.py\n```\n\n### Running the Voice Agent\n\n```bash\npython app/pipeline/voice_agent.py\n```\n\n## Test Your Agent with LiveKit\n\n### 1. Prerequisites Check\n\nMake sure you’ve done this:\n\n1. Activated a LiveKit Cloud instance\n2. Have the following values into .env :\n\n- LIVEKIT_WS_URL=wss://\u003cyour-instance\u003e.livekit.cloud\n- LIVEKIT_API_KEY=...\n- LIVEKIT_API_SECRET=...\n\n### 2. Start Your Voice Agent Locally\n\nRun your `livekit_backend.py` script from terminal:\n\n```bash\npython app/pipeline/livekit_backend.py\n```\n\nIf working correctly, logs will say:\n\nConnected to room `your-livekit-room` as `your-participant-name`\n\nThis means the agent is live and ready to receive audio.\n\n### 3. Join the Same Room as a Human User\n\nUse the LiveKit Agent Playground:\nhttps://agent.livekit.io\n\nThis is essential for testing as the \"other participant\"\n\nSteps:\n\n1. Go to the Playground URL\n2. Input the same Room Name (your-livekit-room)\n3. Use your LiveKit credentials:\n  - API Key, API Secret\n  - Click Join Room\n\n🎙️ Now when you speak, your local VoiceAgentBot should:\n\n1. Detect your voice\n2. Transcribe it\n3. Send it to the LLM\n4. Reply back via audio in real-time\n5. Log metrics\n\n## Performance Metrics\n\nThe system tracks several key metrics:\n- EOU (End of Utterance) Delay\n- TTFT (Time to First Token)\n- TTFB (Time to First Byte)\n- Total Latency\n\n## Future Improvements\n\n1. Smarter language detection\n2. Improved end-of-utterance (EOU) timing\n3. Web or mobile interface integration\n\n## 📜 License\n\nThis project is created for the proPAL AI Backend Engineering Internship assignment.\n\n**Built with ❤️ for proPAL**","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fallwin107%2Fai-voice-agent","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fallwin107%2Fai-voice-agent","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fallwin107%2Fai-voice-agent/lists"}