Metadata-Version: 2.4
Name: trnscrb
Version: 0.1.0
Summary: Offline meeting transcription for macOS — auto-detects meetings, transcribes locally, saves to Claude Desktop via MCP
Author-email: Ajay Ram <ajayrmk@gmail.com>
License: MIT License
        
        Copyright (c) 2025 Ajay Ram
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
        
Project-URL: Homepage, https://github.com/ajayrmk/trnscrb
Project-URL: Repository, https://github.com/ajayrmk/trnscrb
Project-URL: Bug Tracker, https://github.com/ajayrmk/trnscrb/issues
Keywords: transcription,meeting,whisper,macos,mcp,claude,offline
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: MacOS X
Classifier: Intended Audience :: End Users/Desktop
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: MacOS
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
Classifier: Topic :: Office/Business
Requires-Python: >=3.11
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: rumps>=0.4.0
Requires-Dist: sounddevice>=0.4.6
Requires-Dist: numpy>=1.24
Requires-Dist: faster-whisper>=1.0.0
Requires-Dist: pyannote.audio>=3.1
Requires-Dist: mcp>=1.0.0
Requires-Dist: click>=8.1
Requires-Dist: anthropic>=0.25
Requires-Dist: scipy>=1.11
Dynamic: license-file

# trnscrb

Lightweight, fully offline meeting transcription for Claude Desktop.
A native macOS alternative to Granola — no cloud, no subscription.

## What it does

- Auto-detects meetings (Google Meet, Zoom, Slack Huddle, Teams, and more)
- Records your meeting audio (mic + system audio via BlackHole)
- Transcribes locally with Whisper (`small` model, Apple Silicon Metal)
- Identifies speakers with pyannote diarization
- Reads your calendar to auto-name meetings
- Saves every transcript as a plain `.txt` file in `~/meeting-notes/`
- Exposes everything to Claude Desktop via MCP so Claude can search, read, and enrich your notes

## Requirements

- macOS 13+
- Python 3.11+
- Apple Silicon recommended (M1/M2/M3) for fast local transcription

## Install

```bash
# One-liner bootstrap:
bash <(curl -fsSL https://raw.githubusercontent.com/ajayrmk/trnscrb/main/install.sh)

# Or if already cloned:
trnscrb install
```

The installer checks and optionally installs:
- BlackHole 2ch audio driver (via Homebrew)
- Python packages (faster-whisper, pyannote, rumps, sounddevice, mcp, …)
- HuggingFace token (for pyannote speaker diarization)
- Whisper `small` model (~500 MB, downloaded once)
- Claude Desktop MCP config entry
- Launch-at-login agent

## Usage

### Menu bar app
```bash
trnscrb start
```
A mic icon appears in your menu bar with **Auto-transcribe** on by default — it will
automatically start and stop transcribing when it detects a meeting.

Or manually: click → **Start Transcribing** / **Stop Transcribing**.

### From Claude Desktop
Once the MCP server is configured (done by `trnscrb install`), Claude has access to:

| Tool | What it does |
|---|---|
| `start_recording` | Begin capturing audio |
| `stop_recording` | Stop, transcribe, save |
| `recording_status` | Check progress |
| `get_last_transcript` | Most recent transcript |
| `list_transcripts` | All saved meetings |
| `get_transcript` | Full text of one meeting |
| `get_calendar_context` | Current calendar event |
| `enrich_transcript` | Summary + action items via Claude API |

### CLI
```bash
trnscrb list               # list all transcripts
trnscrb show <id>          # print a transcript
trnscrb enrich <id>        # run Claude LLM pass (needs ANTHROPIC_API_KEY)
trnscrb mic-status         # live mic activity monitor (debug)
trnscrb devices            # list audio input devices
```

## System audio (BlackHole setup)

After installing BlackHole:
1. Open **Audio MIDI Setup** (Spotlight → "Audio MIDI Setup")
2. Click **+** → **Create Multi-Output Device**
3. Check both **BlackHole 2ch** and **MacBook Pro Speakers**
4. Go to **System Settings → Sound → Output** → select the Multi-Output Device

## Transcript format

```
Meeting: Weekly Standup
Date:    2024-01-15 10:00
Duration:23:14

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

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

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

After running `trnscrb enrich <id>`, speaker labels are replaced with inferred names
and a summary + action items block is appended.

## Privacy

All audio processing happens on your device. No data leaves your machine except:
- If you run `enrich`, the transcript text is sent to Claude API (Anthropic)
- Calendar access uses AppleScript locally — no network calls

## License

MIT
