local-transcription-mcp--parakeet-tdt-0.6b-v2--

local-transcription-mcp--parakeet-tdt-0.6b-v2--

3
entertainment_and_media

The Parakeet Transcription MCP Server transcribes audio and video files to text using NVIDIA's Parakeet TDT 0.6B V2 model. It supports various audio formats and offers tools to fetch model and system specs, making it effective for high-quality English transcriptions.

Parakeet Transcription MCP Server

Version: 0.1.0

This is an MCP (Model Context Protocol) server designed to transcribe audio and video files into text using NVIDIA's powerful Parakeet TDT 0.6B V2 model. It also offers tools to get details about the model itself.

Quickstart

  • Install Prerequisites: Ensure mise, uv, and FFmpeg are installed.
  • Clone the Repository: Clone the repository and navigate into the project directory.
  • Set up Environment: Use mise to install the correct Python version and activate the environment.
  • Install Dependencies: Use uv to install required Python packages.
  • Run the Server: Start the MCP server using fastmcp.

Features

  • Transcribe audio/video formats.
  • Automatic conversion of input audio.
  • Include detailed word and segment timestamps.
  • Retrieve information about the ASR model and system hardware specifications.

Prerequisites

  • Python: Version 3.12
  • mise: Manage Python versions.
  • uv: Fast Python package installer.
  • FFmpeg: Required for audio/video conversions.