# Whisper: Transcribe longer or multilingual audio locally.

https://thedollscout.com/open-source/whisper
Published by TDS Document Scout
Updated: 2026-09-27

Transcribe longer or multilingual audio locally.

Choose the Python project for its model choices and language coverage. Setup needs Python, FFmpeg and enough resources for the selected model. For a short English clip, the TDS browser tool has a smaller setup.

Transcription can hallucinate, especially with silence or background noise. Check language, names and times against the recording. TDS currently accepts English audio up to 60 seconds.

Install using the upstream setup instructions.

Choose the audio language and a model that fits your hardware.

Test one clip with clear speech before a long job.

Review the transcript and timing against the original.

Python · FFmpeg

MIT

https://github.com/openai/whisper/blob/main/LICENSE

https://github.com/openai/whisper

https://github.com/openai/whisper#setup

https://github.com/openai/whisper

https://thedollscout.com/ai-audio-to-text

A curated directory of independent projects. External software has its own requirements and privacy practices. Listed licenses describe the upstream code; model weights, bundled components and hosted services may have different terms.

Reference material