yazinsai/OpenOats

What it solves

OpenOats is a real-time meeting assistant that helps users stay prepared during calls by transcribing conversations and automatically surfacing relevant information from the user's own personal notes.

How it works

The app runs on macOS and captures system audio and microphone input to provide a live transcript of both speakers. It uses a RAG (Retrieval-Augmented Generation) pipeline to monitor the conversation for key moments—such as questions or decision points—and then searches a local folder of Markdown or text files (the knowledge base) to suggest talking points. Users can choose between fully local processing using Ollama for LLMs and embeddings, or cloud-based providers like OpenRouter and Voyage AI.

Who it’s for

It is designed for professionals who conduct many meetings and want a discreet way to access their research, pitch decks, or customer briefs without manually searching through notes while speaking.

Highlights

• Privacy-First Transcription: Speech recognition is performed entirely offline on the Mac; audio never leaves the device.
• Flexible LLM Integration: Supports local models via Ollama or cloud models via OpenRouter and Requesty.
• Discreet Interface: The app window is hidden from screen sharing by default to remain invisible to other meeting participants.
• Local Knowledge Base: Indexes local .md and .txt files to provide context-aware suggestions during live calls.