trnscrb logo

trnscrb

Paid

Offline meeting transcription for macOS

Inputs: audioOutputs: text
Type
Saas
Company
trnscrb

About trnscrb

trnscrb is a local meeting transcription tool for macOS that lives in the menu bar and automatically detects meetings on platforms like Zoom, Google Meet, Microsoft Teams, Slack, and FaceTime. It utilizes OpenAI's Whisper model to perform on-device transcription via Apple Silicon Metal acceleration, ensuring that no audio or transcript data is sent to the cloud. The tool integrates with Claude Desktop through the Model Context Protocol (MCP), allowing users to search, summarize, and extract action items from their meeting history directly through the Claude interface. It saves transcripts as plain .txt files with speaker diarization and requires no account or subscription.

How to Use

Users can install trnscrb via Homebrew using 'brew install trnscrb' or via Python's pip. After installation and setup (which includes installing the BlackHole audio driver for system audio), the app sits in the macOS menu bar. It automatically begins recording when it detects a meeting app running with an active microphone. Once the meeting ends, it transcribes the audio locally and saves the file to the '~/meeting-notes/' folder, where it can be accessed or queried via Claude Desktop.

Key Features

  • Automatic meeting detection for major platforms (Zoom, Meet, Teams, Slack, FaceTime)
  • On-device transcription using Whisper on Apple Silicon (Metal acceleration)
  • Claude Desktop MCP integration for natural language search and summarization
  • Local-first privacy with no cloud, no accounts, and no subscriptions
  • Speaker diarization for identifying different participants
  • Plain text storage for easy data ownership and portability

Use Cases

  • Transcribing sensitive business meetings without uploading audio to third-party servers
  • Automatically generating action items from a daily standup using Claude
  • Searching across months of meeting history to find specific mentions of a project or person
  • Capturing both microphone and system audio for comprehensive call records

Key Features

Automatic meeting detection for major platforms (Zoom, Meet, Teams, Slack, FaceTime)
On-device transcription using Whisper on Apple Silicon (Metal acceleration)
Claude Desktop MCP integration for natural language search and summarization
Local-first privacy with no cloud, no accounts, and no subscriptions
Speaker diarization for identifying different participants
Plain text storage for easy data ownership and portability

Pros & Cons

Pros
  • Complete privacy with fully on-device transcription
  • Free to use with no subscription or account needed
  • Automatic platform detection reduces manual setup
  • High performance via Apple Silicon Metal acceleration
  • Seamless Claude Desktop integration for advanced analysis
  • Simple .txt output with speaker labels for easy sharing and archiving
Cons
  • Exclusive to macOS and requires Apple Silicon
  • Advanced features like summarization require Claude Desktop
  • No built-in cloud sync or multi-device support
  • Limited to supported meeting platforms only
  • Potential setup needed for optimal Whisper model performance

Best For

Transcribing sensitive business meetings without uploading audio to third-party serversAutomatically generating action items from a daily standup using ClaudeSearching across months of meeting history to find specific mentions of a project or personCapturing both microphone and system audio for comprehensive call records

Alternatives to trnscrb

FAQ

Is trnscrb free to use?
Yes, it requires no account or subscription and is available via GitHub.
Which meeting platforms does it support?
It automatically detects and transcribes from Zoom, Google Meet, Microsoft Teams, Slack, and FaceTime.
Does it send data to the cloud?
No, all transcription happens on-device using OpenAI's Whisper with Apple Silicon Metal acceleration.
What output format does it produce?
Transcripts are saved as plain .txt files with speaker diarization.
How does it integrate with Claude?
Via the Model Context Protocol (MCP), allowing search, summarization, and action item extraction in Claude Desktop.
What hardware is required?
macOS on Apple Silicon for Metal acceleration.