mjaskolski / AIVoiceChat

Low latency ai companion voice talk in 60 lines of code using faster_whisper and elevenlabs input streaming

Seamless and real-time voice interaction with AI.

Uses faster_whisper and elevenlabs input streaming for low latency responses to spoken input.

🎥 Watch a Demo Video

Note: The demo is conducted on a 10Mbit/s connection, so actual performance might be more impressive on faster connections.

voice_talk_vad.py - automatically detects speech

voice_talk.py - toggle recording on/off with the spacebar

🛠 Setup:

1. API Keys:

Replace your_openai_key and your_elevenlabs_key with your OpenAI and ElevenLabs API key values in the code.

2. Dependencies:

Install the required Python libraries:

pip install openai elevenlabs pyaudio wave keyboard faster_whisper numpy torch 

3. Run the Script:

Execute the main script based on your mode preference:

python voice_talk_vad.py

or

python voice_talk.py

🎙 How to Use:

For voice_talk_vad.py:

Talk into your microphone.
Listen to the reply.

For voice_talk.py:

1. Press the space bar to initiate talk.
2. Speak your heart out.
3. Hit the space bar again once you're done.
4. Listen to reply.

🤝 Contribute

Feel free to fork, improve, and submit pull requests. If you're considering significant changes or additions, please start by opening an issue.

💖 Acknowledgements

Huge shoutout to:

• The hardworking developers behind faster_whisper.
• ElevenLabs for their cutting-edge voice API.
• OpenAI for pioneering with the GPT-4 model.