Skip to main content

trnscrb

Offline meeting transcription for macOS — no cloud, no subscription.

trnscrb lives in your menu bar, listens for meetings, transcribes them locally with Whisper, and makes every transcript searchable from Claude Desktop via MCP.


Install

brew tap ajayrmk/tap
brew install trnscrb
trnscrb install

Or with pip / uv:

pip install trnscrb && trnscrb install
uv tool install trnscrb && trnscrb install

trnscrb install is a guided setup that handles:

  • BlackHole 2ch audio driver (captures system audio alongside mic)
  • HuggingFace token for speaker diarization (pyannote)
  • Whisper small model download (~500 MB, one-time)
  • Claude Desktop MCP config
  • Launch-at-login agent

Quick start

trnscrb start       # launch the menu bar app

With Auto-transcribe on (the default), trnscrb detects when a meeting starts — Google Meet, Zoom, Slack Huddle, Teams, FaceTime — and begins recording automatically. When the meeting ends, it stops, transcribes, and saves.

You can also trigger manually from the menu bar: Start Transcribing / Stop Transcribing.


How it works

Step What happens
Meeting detected Mic active for 5 s + meeting app found
Recording Audio captured via mic or BlackHole (system + mic)
Transcription Whisper small model, runs locally on Apple Silicon
Diarization Speaker labels via pyannote (needs HuggingFace token)
Saved Plain .txt in ~/meeting-notes/

Claude Desktop integration

After trnscrb install, Claude Desktop has these tools available:

Tool Description
start_recording Start capturing audio
stop_recording Stop and transcribe in the background
recording_status Check if recording or transcribing
get_last_transcript Fetch the most recent transcript
list_transcripts List all saved meetings
get_transcript Read a specific transcript
get_calendar_context Current or upcoming calendar event
enrich_transcript Add summary + action items via Claude API

CLI

trnscrb start               # launch menu bar app
trnscrb install             # guided setup / re-check dependencies
trnscrb list                # list saved transcripts
trnscrb show <id>           # print a transcript
trnscrb enrich <id>         # summarise + action items (needs ANTHROPIC_API_KEY)
trnscrb mic-status          # live mic activity monitor — useful for debugging
trnscrb devices             # list audio input devices
trnscrb watch               # headless auto-transcribe, no menu bar

System audio with BlackHole

To capture both your mic and the other participants' audio:

  1. Install BlackHole via trnscrb install (or brew install blackhole-2ch)
  2. Open Audio MIDI Setup → + → Create Multi-Output Device
  3. Check BlackHole 2ch and MacBook Pro Speakers
  4. System Settings → Sound → Output → select the Multi-Output Device

trnscrb auto-detects BlackHole and uses it when available. Without it, only your mic is recorded.


Transcript format

Meeting: Weekly Standup
Date:    2025-02-18 10:00
Duration:23:14

============================================================

[SPEAKER_00]
  00:12  Good morning, let's get started.

[SPEAKER_01]
  00:18  Morning! I finished the auth PR yesterday.

Running trnscrb enrich <id> replaces SPEAKER_00 / SPEAKER_01 with inferred names and appends a summary and action items block.


Requirements

  • macOS 13 or later
  • Python 3.11+
  • Apple Silicon (M1/M2/M3/M4) recommended — Whisper runs on Metal

Privacy

Everything runs on your machine. No audio or transcripts leave your device unless you explicitly run enrich, which sends the transcript text to the Claude API.


License

MIT

Metadata

Release files for trnscrb 0.1.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for trnscrb 0.1.2
File Size Uploaded Provenance Actions
trnscrb-0.1.2.tar.gz 25.7 kB Without provenance Details for trnscrb-0.1.2.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for trnscrb 0.1.2
File Interpreter ABI Platform Actions
trnscrb-0.1.2-py3-none-any.whl Python 3 none any Details for trnscrb-0.1.2-py3-none-any.whl

Total release size: 54.4 kB

Release files / trnscrb-0.1.2.tar.gz

Download URL trnscrb-0.1.2.tar.gz
Size 25.7 kB
Tags Source
SHA-256 checksum
How to use checksums
8ec52014ede1a6f1aecabacd33adbc350b6fc07a30359fc30c2b3dc587e90ac2
BLAKE2b-256 checksum
How to use checksums
d548640f8f63286cd80ac9b453ee491f134e984c54e4d52a7f1406861b4364fa
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via uv/0.10.4 {"installer":{"name":"uv","version":"0.10.4","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"macOS","version":null,"id":null,"libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

Release files / trnscrb-0.1.2-py3-none-any.whl

Download URL trnscrb-0.1.2-py3-none-any.whl
Size 28.7 kB
Tags Python 3
SHA-256 checksum
How to use checksums
e51bb246933bc69c72a1b8125f0653c422049e0d47d83047684e9de1ced78f07
BLAKE2b-256 checksum
How to use checksums
c39731350dc1a00511965b4a8a541dd4abd23921f27fa2e3c635ddbfaa75cc29
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via uv/0.10.4 {"installer":{"name":"uv","version":"0.10.4","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"macOS","version":null,"id":null,"libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

Security for trnscrb 0.1.2

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page