DISCOVER THE FUTURE OF AI AGENTS

Minutes

Added May 7, 2026
Agent & Tooling
Open Source
RustKnowledge BaseModel Context ProtocolRAGAI AgentsCLIAgent & ToolingDocs, Tutorials & ResourcesKnowledge Management, Retrieval & RAGEnterprise Applications & Office

Open-source, local-first conversation memory infrastructure that transcribes meetings and voice memos into structured Markdown, queryable by humans and AI agents.

Minutes is an open-source conversation memory tool built with Rust + Tauri by Mat Silverstein. Core capabilities span local recording and multi-engine transcription (whisper.cpp / Parakeet / Apple Speech), pyannote-rs local diarization, optional LLM summarization and structured extraction (Key Points / Decisions / Action Items / Commitments, etc.), and SQLite-based cross-meeting relationship graph queries.

Output is Markdown files with YAML frontmatter stored locally in ~/meetings/, compatible with Obsidian / Logseq / grep / git, ensuring long-term data accessibility.

For AI agent integration, it provides 29 MCP tools (launched via npx minutes-mcp), natively supporting Claude Desktop, Claude Code, Codex, OpenCode, Gemini CLI, Mistral Vibe and other clients, enabling cross-meeting contextual queries (e.g., "What did I commit to Sarah?"), live meeting coaching, and more.

Additional features include: real-time transcription stream (JSONL), dictation mode, voice memo auto-processing pipeline (minutes watch), template system (4 built-in + layered overrides), 50+ CLI commands, command palette (⌘⇧K), Granola history import, and more. The desktop app runs as a menu bar app on macOS and offers a standalone installer on Windows. MIT licensed, current version v0.16.4.

Core Pipeline#

Audio → Transcribe → Diarize → Summarize → Structured Markdown → Relationship Graph
         (local)     (local)    (LLM)       (decisions,            (people, commitments,
        whisper.cpp  pyannote   Claude/       action items,          topics, scores)
        /parakeet    -rs        Ollama/       people, entities)      SQLite index
                                Mistral/OpenAI

Installation#

  • macOS desktop app: brew install --cask silverstein/tap/minutes
  • macOS CLI only: brew tap silverstein/tap && brew install minutes
  • Windows: Download .exe installer from GitHub Releases
  • Cross-platform from source: cargo install minutes-cli
  • MCP server (no Rust needed): npx minutes-mcp
  • Initial setup: minutes setup --model small to download Whisper model (466MB)
  • Quick demo: minutes demo --full to seed 5 sample meetings

Related Projects

View All

STAY UPDATED

Get the latest AI tools and trends delivered straight to your inbox. No spam, just intelligence.