You can create a subtitle draft without sending your video to a transcription server. Whisper AI’s free local workflow reads the media in your browser, recognizes the speech and produces editable timed cues. The result is an SRT or VTT subtitle file; it does not render a new video with captions burned into the picture.

Start with the final video edit

Use the version you intend to publish. Cutting an introduction or rearranging a scene after generating subtitles can leave the captions out of sync. Local mode accepts files up to 200 MB and 20 minutes. Video decoding depends on your browser and the codec inside the file.

If the video cannot be decoded, try a supported audio export from the same final edit. Keep the same start point and duration so its timings still match the video. For longer media, split it carefully or review the separate cloud transcription option.

Generate the first subtitle draft

  1. Open the AI subtitle generator and choose Upload a file.
  2. Select your video and set Speech language to its spoken language, or use automatic detection.
  3. Under Advanced settings, start with Base and Compatibility mode. Try Tiny if your device struggles.
  4. Choose Start transcription and leave the page open while the model downloads and processes the recording.

The first run needs a model download. That network request does not upload the video. Using the separate cloud workflow does upload the selected media for processing, so choose the mode that fits your needs before starting.

Edit wording and timing together

Listen to the recording while checking each cue. Correct names, punctuation, technical terms and missing words. Use the cue’s Start and End fields to adjust its display interval. You can add a missing cue or delete an incorrect one; the editor also offers undo for a deletion.

Check that each cue ends after it starts and that lines do not overlap unintentionally. Listen around pauses and changes of speaker. Recognition timestamps are a starting point, not a guarantee of frame-perfect synchronization.

For accessible captions, review meaningful non-speech sounds and speaker identification yourself. Automatic speech transcription does not reliably add all of the context a viewer may need.

Export and test in the destination

In Export format, choose SRT or VTT and select Export. Pick the format your destination accepts; the SRT vs VTT guide explains the differences. You can change formats later with the subtitle converter.

Import the subtitle file alongside your video in the target editor or player. Check its opening cue, a middle section and the ending. If you want permanent text in the image, use a video editor to burn in the reviewed subtitles and render a new video.

Keep a copy outside the browser

Local history stays in this browser on this device and can disappear when browser storage is cleared. Save the exported subtitle file with the source video. Keep a separate transcript export if you also need readable notes, and review the user guide before relying on browser history as your only copy.