minutes

Open-source conversation memory for AI assistants. Record meetings, capture voice memos, search everything — your AI remembers every conversation you've had.

Website | GitHub | MCP Server | Desktop App

Install

cargo install minutes-cli
minutes setup --model tiny    # Download whisper model (75MB)

Or via Homebrew on macOS:

brew tap silverstein/tap && brew install minutes

Quick start

minutes record                # Start recording
minutes stop                  # Stop and transcribe
minutes search "pricing"      # Search across all meetings
minutes actions               # Open action items

What it does

• Record meetings and voice memos from your microphone
• Transcribe locally with whisper.cpp (Apple Silicon optimized)
• Diarize speakers with pyannote-rs (native Rust, no Python)
• Extract action items, decisions, and people into structured YAML frontmatter
• Search across all conversations by keyword, person, or topic
• Track commitments and relationships across meetings
• Sync voice memos from your phone via iCloud/Dropbox/any folder sync
• Integrate with Claude, Cursor, Windsurf, Obsidian via MCP

Key commands

Output format

Meetings save as markdown with structured YAML frontmatter:

---
title: Q2 Pricing Discussion
date: 2026-03-17T14:00:00
duration: 42m
action_items:
  - assignee: mat
    task: Send pricing doc
    due: Friday
    status: open
decisions:
  - text: Run pricing experiment with monthly billing
---

Works with Obsidian, Logseq, grep, or any markdown tool.

Features

Claude / MCP integration

No API keys needed — Claude reads your meetings via MCP tools.

# MCP server (no Rust required)
npx minutes-mcp

You: "What did Alex say about pricing?"
Claude: [searches meetings] → synthesizes answer from transcripts

GPU acceleration

Metal is the only backend that is exercised daily by the maintainer. CUDA, ROCm/HIP, and Vulkan should be considered experimental: they wire through to whisper.cpp via whisper-rs and are expected to work, but have not been validated in CI.

cargo install minutes-cli --features metal    # macOS Metal
cargo install minutes-cli --features coreml   # macOS Neural Engine
cargo install minutes-cli --features cuda     # NVIDIA CUDA
cargo install minutes-cli --features hipblas  # AMD ROCm/HIP (experimental)
cargo install minutes-cli --features vulkan   # Vulkan (experimental)

Windows CUDA users: You may need to set environment variables before building:

$env:CUDA_PATH = "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4"
$env:CMAKE_CUDA_COMPILER = "$env:CUDA_PATH\bin\nvcc.exe"
$env:LIBCLANG_PATH = "C:\Program Files\LLVM\bin"
$env:CMAKE_GENERATOR = "NMake Makefiles"

The first CUDA build takes longer than usual (compiling GPU kernels) — this is a one-time cost.

ROCm/HIP users: The build expects ROCm installed at /opt/rocm. If your installation is elsewhere, set HIP_PATH before building:

export HIP_PATH=/path/to/rocm

Vulkan users: On Windows and macOS, set VULKAN_SDK to your SDK install root before building. On Linux, whisper-rs-sys links against the system libvulkan.

Links

• Website: useminutes.app
• GitHub: github.com/silverstein/minutes
• MCP Server: npmjs.com/package/minutes-mcp
• Desktop App: GitHub Releases
• Claude Code Plugin: claude plugin marketplace add silverstein/minutes

License

MIT — Built by Mat Silverstein, founder of X1 Wealth