minutes

Your AI remembers every conversation you've had.

Agents have run logs. Humans have conversations. minutes captures the human side — the decisions, the intent, the context that agents need but can't observe — and makes it queryable.

Record a meeting. Capture a voice memo on a walk. Ask Claude what was decided three weeks ago. It just works.

$ minutes record --context "Q2 pricing discussion with Alex"
Recording... (Ctrl-C or `minutes stop` to finish)
  Tip: add notes with `minutes note "your note"` in another terminal

^C
Stopping recording...
Transcribing.....
Saved: ~/meetings/2026-03-17-q2-pricing-discussion-with-sarah.md

$ minutes search "pricing"
2026-03-17 — Q2 Pricing Discussion with Alex [meeting]
  [4:20] I think monthly billing makes more sense for independent advisors...

$ minutes actions --assignee mat
Open action items (2):
  @user: Send pricing doc (by Friday)
    from: 2026-03-17 — Q2 Pricing Discussion with Alex
  @user: Set up monthly billing tier experiment (by next week)
    from: 2026-03-17 — Q2 Pricing Discussion with Alex

How it works

Audio → Transcribe → Summarize → Detect Attendees → Structured Markdown
         (local)       (LLM)      (calendar +          (decisions,
        whisper.cpp   Claude/      transcript)          action items,
                      Ollama/                           people, entities)
                      OpenAI

Everything runs locally. Your audio never leaves your machine (unless you opt into cloud LLM summarization). Attendees are automatically detected from macOS Calendar events and participant extraction from the transcript.

Install

macOS

# Homebrew
brew tap silverstein/tap
brew install minutes

# Or from source (requires Rust + cmake)
export CXXFLAGS="-I$(xcrun --show-sdk-path)/usr/include/c++/v1"
cargo install --path crates/cli

Windows

# Download pre-built binary from GitHub releases, or build from source:
# Requires: Rust, cmake, MSVC build tools
cargo install --path crates/cli

Linux

# Requires: Rust, cmake, ALSA dev headers
sudo apt-get install -y libasound2-dev  # Debian/Ubuntu
cargo install --path crates/cli

Setup (all platforms)

# Download whisper model
minutes setup --model tiny    # Quick start (75MB, fast, less accurate)
minutes setup --model small   # Recommended (466MB, good accuracy)
minutes setup --model base    # Middle ground (141MB)

Platform notes: Calendar integration (auto-detecting meeting attendees) requires macOS. Screen context capture works on macOS and Linux. All other features — recording, transcription, search, action items, person profiles — work on all platforms.

Desktop app (optional)

export CXXFLAGS="-I$(xcrun --show-sdk-path)/usr/include/c++/v1"
export MACOSX_DEPLOYMENT_TARGET=11.0
cargo tauri build --bundles app
# Opens: target/release/bundle/macos/Minutes.app

The desktop app adds a system tray icon, recording controls, audio visualizer, and a meeting list window. macOS will prompt for microphone permission on first recording.

Privacy: All Minutes windows are hidden from screen sharing by default — other participants on Zoom/Meet/Teams won't see the app. Toggle via the tray menu: "Hide from Screen Share ✓".

Signed macOS releases

Maintainers can build signed and notarized desktop releases with the GitHub Actions workflow in .github/workflows/release-macos.yml.

The full secrets contract and maintainer steps live in docs/RELEASE-MACOS.md.

Stable vs preview channels and rollback expectations live in docs/RELEASE-CHANNELS.md. Reproducible release-note generation lives in docs/RELEASE-NOTES.md. The current auto-update evaluation and recommendation live in docs/AUTO-UPDATE-EVALUATION.md.

Troubleshooting

No speech detected / blank audio: The most common cause is microphone permissions. Check System Settings → Privacy & Security → Microphone and ensure your terminal app (or Minutes.app) has access.

tmux users: tmux server runs as a separate process that doesn't inherit your terminal's mic permission. Either run minutes record from a direct terminal window (not inside tmux), or use the Minutes.app desktop bundle which gets its own mic permission.

Build fails with C++ errors on macOS 26+: whisper.cpp needs the SDK include path. Set CXXFLAGS as shown above before building.

Features

Record meetings

minutes record                                    # Record from mic
minutes record --title "Standup" --context "Sprint 4 blockers"  # With context
minutes stop                                      # Stop from another terminal

Take notes during meetings

minutes note "Alex wants monthly billing not annual billing"          # Timestamped, feeds into summary
minutes note "Logan agreed"                       # LLM weights your notes heavily

Process voice memos

minutes process ~/Downloads/voice-memo.m4a        # Any audio format
minutes watch                                     # Auto-process new files in inbox

Search everything

minutes search "pricing"                          # Full-text search
minutes search "onboarding" -t memo               # Filter by type
minutes actions                                   # Open action items across all meetings
minutes actions --assignee sarah                   # Filter by person
minutes list                                      # Recent recordings

Post-meeting annotations

minutes note --meeting ~/meetings/2026-03-17-pricing.md "Alex confirmed via email"

Screen context (opt-in)

Capture periodic screenshots during recording so the LLM knows what was on screen when someone said "look at this number." Screenshots are sent as images to vision-capable LLMs (Claude, GPT-4o).

# ~/.config/minutes/config.toml
[screen_context]
enabled = true
interval_secs = 30

Privacy: disabled by default, screenshots are 0600 permissions, auto-deleted after summarization, downscaled to 1280px. Requires Screen Recording permission on macOS.

Output format

Meetings save as markdown with structured YAML frontmatter:

---
title: Q2 Pricing Discussion with Alex
type: meeting
date: 2026-03-17T14:00:00
duration: 42m
context: "Discuss Q2 pricing, follow up on annual billing decision"
action_items:
  - assignee: mat
    task: Send pricing doc
    due: Friday
    status: open
  - assignee: sarah
    task: Review competitor grid
    due: March 21
    status: open
decisions:
  - text: Run pricing experiment at monthly billing with 10 advisors
    topic: pricing experiment
---

## Summary
- Alex proposed lowering API launch timeline from annual billing to monthly billing/mo
- Compromise: run experiment with 10 advisors at monthly billing

## Notes
- [4:23] Alex wants monthly billing not annual billing
- [12:10] Logan agreed

## Decisions
- [x] Run pricing experiment at monthly billing with 10 advisors

## Action Items
- [ ] @user: Send pricing doc by Friday
- [ ] @sarah: Review competitor grid by March 21

## Transcript
[SPEAKER_0 0:00] So let's talk about the pricing...
[SPEAKER_1 4:20] I think monthly billing makes more sense...

Works with Obsidian, grep, or any markdown tool. Action items and decisions are queryable via the CLI and MCP tools.

Claude integration

minutes is a native extension for the Claude ecosystem. No API keys needed — Claude summarizes your meetings when you ask, using your existing Claude subscription.

The pipeline captures and transcribes locally. Claude does the intelligence:

You: "Summarize my last meeting"
Claude: [calls get_meeting] → reads transcript → summarizes in conversation

You: "What did Alex say about pricing?"
Claude: [calls search_meetings] → finds matches → synthesizes answer

You: "Any open action items for me?"
Claude: [calls list_meetings] → scans frontmatter → reports open items

No ANTHROPIC_API_KEY. No extra cost. Just your Claude subscription doing what it already does — but with your meeting history as context.

Any MCP client (Claude Desktop, Cursor, Windsurf, your own agent)

Minutes exposes a standard MCP server. Point any MCP-compatible client at it:

{
  "mcpServers": {
    "minutes": {
      "command": "node",
      "args": ["path/to/minutes/crates/mcp/dist/index.js"]
    }
  }
}

12 tools: start_recording, stop_recording, get_status, list_meetings, search_meetings, get_meeting, process_audio, add_note, consistency_report, get_person_profile, research_topic, qmd_collection_status

6 resources: minutes://meetings/recent, minutes://status, minutes://actions/open, minutes://events/recent, minutes://meetings/{slug}, ui://minutes/dashboard

Interactive dashboard (Claude Desktop): 5 tools render an inline interactive UI via MCP Apps — meeting list with filter/search, detail view with fullscreen + "Send to Claude" context injection, person profiles, consistency reports. Text-only clients see the same data as plain text.

Claude Code (Plugin)

.claude/plugins/minutes/
├── 11 skills:
│   ├── Core: /minutes record, search, list, note, verify, setup, cleanup
│   └── Interactive: /minutes prep, debrief, recap, weekly
├── 1 agent: meeting-analyst (cross-meeting intelligence)
└── 2 hooks: post-recording alerts + proactive meeting reminders

Meeting lifecycle skills — inspired by gstack's interactive skill pattern:

/minutes prep "call with Alex"     → relationship brief, talking points, .prep.md saved
  ↓
minutes record → minutes stop       → hook alerts if decisions conflict with prior meetings
  ↓
/minutes debrief                    → "You wanted to resolve pricing. Did you?"
  ↓
/minutes weekly                     → themes, decision arcs, stale items, Monday brief

Each interactive skill pushes back on vague input, produces actionable intelligence, and ends with a three-beat closing ritual (signal reflection, concrete assignment, next-skill nudge).

Minutes Desktop Assistant

The Tauri menu bar app includes a built-in AI Assistant window backed by the same local meeting artifacts. It runs as a singleton assistant session:

• AI Assistant opens or focuses the persistent assistant window
• Discuss with AI reuses that same assistant and switches its active meeting focus
• General assistant instructions live in CLAUDE.md; meeting-specific focus is written to CURRENT_MEETING.md

Cowork / Dispatch

MCP tools are automatically available in Cowork. From your phone via Dispatch: "Start recording" → Mac captures → Claude processes → summary on your phone.

Optional: automated summarization

If you want summaries generated automatically in the pipeline (without asking Claude each time):

# Use your existing Claude Code or Codex subscription (recommended)
[summarization]
engine = "agent"
agent_command = "claude"  # or "codex" for OpenAI Codex users

# Or use a free local LLM
[summarization]
engine = "ollama"
ollama_model = "llama3.2"

The agent engine shells out to your installed AI CLI — no API keys to manage, uses the same subscription you already have. Most users won't need this since Claude handles summarization conversationally via MCP.

Import from Granola

Already using Granola? Import your meeting history:

# Preview what will be imported
minutes import granola --dry-run

# Import all Granola meetings to ~/meetings/
minutes import granola

Reads from ~/.granola-archivist/output/ (the default Granola export location). Meetings are converted to Minutes' markdown format with YAML frontmatter. Duplicates are skipped automatically.

Voice memos (iPhone → Mac)

No iOS app needed. Use Apple's built-in Voice Memos + a Shortcut:

1. Run minutes watch on your Mac (or install the launchd service for auto-start)
2. Record in Voice Memos on iPhone
3. Share → "Save to Minutes" (iCloud Drive sync)
4. Markdown appears in ~/meetings/memos/

Supports .m4a, .mp3, .wav, .ogg, .webm. Format conversion is automatic via symphonia.

Configuration

Optional — minutes works out of the box.

# ~/.config/minutes/config.toml

[transcription]
model = "small"           # tiny (75MB), base, small (466MB), medium, large-v3 (3.1GB)

[summarization]
engine = "none"           # Default: Claude summarizes conversationally via MCP
                          # "agent" = uses your Claude Code or Codex subscription (no API key)
                          # "ollama" = local, free
                          # "claude" / "openai" = direct API key (legacy)
agent_command = "claude"  # Which CLI to use when engine = "agent" (claude, codex, etc.)
ollama_url = "http://localhost:11434"
ollama_model = "llama3.2"

[diarization]
engine = "none"           # "none" (default) or "pyannote" (requires pip install pyannote.audio + HuggingFace auth)

[search]
engine = "builtin"        # builtin (regex) or qmd (semantic)

[watch]
paths = ["~/.minutes/inbox"]
settle_delay_ms = 2000    # iCloud sync safety delay

[screen_context]
enabled = false           # Opt-in: capture screenshots during recording for LLM context
interval_secs = 30        # How often to capture (seconds)
keep_after_summary = false # Delete screenshots after summarization (default: clean up)

[assistant]
agent = "claude"          # CLI launched by the Tauri AI Assistant
agent_args = []           # Optional extra args, e.g. ["--dangerously-skip-permissions"]

Architecture

minutes/
├── crates/core/    17 Rust modules — the engine (shared by all interfaces)
├── crates/cli/     CLI binary — 15 commands
├── crates/reader/  Lightweight read-only meeting parser (no audio deps)
├── crates/mcp/     MCP server — 12 tools + 6 resources + interactive dashboard
│   └── ui/         MCP App dashboard (vanilla TS → single-file HTML)
├── tauri/          Menu bar app — system tray, recording UI, singleton AI Assistant
└── .claude/plugins/minutes/   Claude Code plugin — 11 skills + 1 agent + 2 hooks

Single minutes-core library shared by CLI, MCP server, and Tauri app. Zero code duplication.

Building your own agent on Minutes

Minutes is designed as infrastructure for AI agents. The MCP server is the primary integration surface:

• Read meetings: list_meetings, search_meetings, get_meeting return structured JSON
• Track people: get_person_profile builds cross-meeting profiles with topics, open commitments
• Monitor consistency: consistency_report flags conflicting decisions and stale commitments
• Record + process: start_recording, stop_recording, process_audio for pipeline control
• Resources: Stable URIs (minutes://meetings/recent, minutes://actions/open) for agent context injection

Any agent framework that speaks MCP can use Minutes as its conversation memory layer — the agent handles the intelligence, Minutes handles the recall.

Built with: Rust, whisper.cpp, symphonia, cpal, Tauri v2, ureq

Contributing

See CONTRIBUTING.md.

License

MIT