whisper.cpp

whisper.cpp

Use Whisper locally to turn recordings into text or subtitles with AI, then review and organize the result yourself.

Visit website

What is whisper.cpp?

Have an interview, lesson recording or voice note you do not want to type out from scratch? Let whisper.cpp produce a first transcript. It runs Whisper speech-recognition models on your own device and turns recorded speech into text you can edit.

For an interview, you can start with the transcript and correct names and specialist terms. For a video, time-aligned output can help prepare subtitles before you check the wording and timing. It recognizes what was said; summaries and action items can be handled by tools later in the workflow.

Once the model and environment are ready, recognition can happen locally without uploading every recording to a hosted service. Developers can also integrate it into their applications to use speech as input instead of building a separate recognition service.

Command-line tools and browser examples suit users willing to configure software and development teams. Models differ in memory needs and processing time. Noise, accents and specialist terms also affect results, so listen back to the original audio when checking important passages.

加载中 / Loading

Discussion

Loading comments…