MeetingMind - AI Meeting Intelligence & MCP Pipeline

MeetingMind - AI Meeting Intelligence & MCP Pipeline

MeetingMind is an automated AI meeting intelligence pipeline engineered around a multi-tool FastMCP (Model Context Protocol) server integrated with a unified LiteLLM client. It automates post-meeting workflows by turning raw audio or meeting transcripts into structured action items, sentiment profiles, decision logs, and synced tasks.

Core Capabilities

• **Action Item Extraction**: Automatically identifies who needs to complete specific tasks, assigning owners and deadlines.
• **Sentiment Analysis & Blocker Detection**: Evaluates team emotional tone and surfaces critical blockers across discussion topics.
• **Decision Log**: Maintains a structured log of all architectural, product, and business decisions.
• **Multi-Platform Task Integration**: Automatically creates and syncs task items directly to Notion (`notion-client`) and Trello (`py-trello`).
• **Recurring Topic Tracker**: Uses persistent SQLite cross-session memory to detect patterns and recurring blockers across historical meetings.

Tech Stack

• **MCP Framework**: FastMCP (Python)
• **LLM Gateway**: LiteLLM (supporting Ollama, Gemini, OpenAI)
• **Audio Transcription**: OpenAI Whisper
• **Integrations**: Notion Client (`notion-client`), Trello (`py-trello`)
• **Memory / History**: SQLite3
• **CLI Interface**: Typer