theaiautomators/insights-lm-local-package

Supports local audio transcription via Whisper ASR and text-to-speech synthesis through Coqui TTS, enabling fully offline podcast generation and multi-modal document interaction. The backend re-engineers N8N workflows to orchestrate local inference pipelines—LLM embeddings, transcription, and synthesis—while Supabase edge functions handle document processing and citation retrieval without external API calls. Containerized via Docker with GPU acceleration support, the stack decouples from cloud dependencies while maintaining the original frontend's React/Vite interface.

200 stars. No commits in the last 6 months.