GPT Voice Assistant

Interact with GPT via voice as a CLI tool. Context aware assistant via pgvector and text embeddings.

Assistant Layer

OpenAI's Chat Completion API for text and AWS Polly API for speech synthesis.

Storage Layer

PostgreSQL support for storing chat interactions as messages with embeddings. Options to store audio files.

Audio Layer

Uses PyAudio for buffer reading and PyTorch for loading data into Whisper. Uses ffmpeg for playback.

Example