finnvoor/yap

🗣️ A CLI for on-device speech transcription using Speech.framework on macOS 26

What it solves

yap is a command-line interface (CLI) tool that provides easy access to on-device speech transcription. It allows users to convert audio and video files, live system audio, or microphone input into text, avoiding the need for cloud-based transcription services.

How it works

The tool leverages Apple's Speech.framework on macOS to perform transcription on the device. It supports multiple input sources: files (audio/video), live system audio (via Screen Recording permissions), and microphone input (via Microphone permissions).

Who it’s for

Developers and power users on macOS who need a fast, local transcription tool that can be piped into other CLI tools (like LLMs for summarization) or used to generate caption files for video production.

Highlights

  • Multiple Input Modes: Supports file transcription, live system audio capture, and microphone dictation.
  • Flexible Output Formats: Exports to plain text, SRT, VTT, and JSON.
  • Meeting Transcription: The listen-and-dictate command can transcribe both system audio and microphone input simultaneously with custom speaker labels.
  • MCP Server: Includes a Model Context Protocol (MCP) server, allowing AI agents to use the transcribe tool.
  • Local Processing: All transcription is performed on-device using macOS frameworks.

Related

  • Project
  • Project
  • Project
  • Project
  • Project