GitHub radar

Discover who said what and when, without the need for a GPU.

Automatically detects all speakers in audio files and timestamps them. Once you have it installed, you’ll be able to easily create a speaker timeline for your podcast or video.

01pyannote/speaker-diarization-community-1 2.4k5503k downloads/mospeaker-diarization

The program listens to an audio file and marks when each participant is speaking. It doesn't recognize words or write text — only who spoke and when.

When it helps

Useful if you're making a podcast, video, or reviewing a multi-person call and want to quickly see who spoke when. Skip it if your recording has one voice or you need an actual transcript.

Pros

  • Runs on a regular computer, a GPU isn't required
  • Works fully offline, no internet connection needed
  • Fewer errors than the previous version of the model
  • You can tell it the number of speakers if known

Cons

  • Model is gated — accept terms and get a Hugging Face token first
  • Doesn't recognize words itself — only identifies speakers

How to set it up — step by step

  1. 1Open your AI agent (Claude Code, Codex) in your project folder
  2. 2Prepare the audio file with multiple speakers you want to process
  3. 3Send the agent the text from the block below
  4. 4Give the agent the file — it will set up a Hugging Face token and install the pyannote.audio library
  5. 5Check the result: the agent should show who spoke and when in the recording

Text for your agent

Copy this and send it to your agent — Claude Code, Codex, any of them:

Set up Pyannote's speaker-labeling model for me: https://huggingface.co/pyannote/speaker-diarization-community-1 - get a Hugging Face token, install pyannote.audio, and run it on this recording (I'll give you the file), showing who spoke when.

Runs on a regular computer without a GPU - it processes on the CPU by default, a GPU just speeds it up. You need Python and the pyannote.audio library.

Open on Hugging Face