AI Blocks › Turn speech into text and back
Whisper speech transcriber
Give it an audio file. Get a written transcript in 99 languages, or an English translation.
Whisper speech transcriber is a free AI skill for the job "Turn speech into text and back". Give the instructions to an AI assistant that supports the required files and tools, then review its output.
Free · No account · MIT licence · Made by Orchestra Research · Level: beginner · Thinking: low
Made by Orchestra Research. See the original on GitHub; we copied it as it is, with its licence.
What it does
A written transcript of your audio, in the original language or translated into English. Podcasts, interviews and meeting recordings become searchable text. It runs Whisper, a speech model trained on 680,000 hours of audio and covering 99 languages. It copes well with noise and accents. The skill covers installing the model, choosing a model size that balances speed against accuracy, running transcription and translation, and saving the result with timestamps. It also says when another tool fits better, such as a managed service if you need speaker labels or live streaming. You need Python and a computer able to run the model.
You give it
- An audio or video file
- Language, if you know it (optional)
- Whether you want a translation
- Computer with Python
You get
- Full text transcript
- Timestamps
- English translation (optional)
- Choice of model size
- Saved transcript file
Use it when
- You have a recorded interview and need the words on a page.
- Your podcast needs show notes or subtitles.
- A recording is in another language and you want it in English.
- Meeting audio is noisy and other tools mishear it.
How to use it
- Add it to your AI once: download it and add the folder to Claude, ChatGPT or Codex as a skill.
- Ask for the job in your own words. Your AI opens the skill when it fits.
- It works best when your AI can run Python code.