Original media
Local playbackHow to use the editor
- Open an audio or video file. The app reads it on your device and builds its waveform.
- Choose Tiny English for a smaller download, or Base English for a larger model. Set a maximum number of characters and select Transcribe English. Keep the tab open and your device awake. Speed and memory needs depend on your device.
- Check the transcript. Names, accents, background music, silence and overlapping voices can cause mistakes. Timing is an estimate and also needs checking.
- Edit the text under the player. Select a block in the timeline, drag either edge, or type its In and Out times. Set in and Set out use the playhead. Play block plays just that block.
- To split a block, put the text cursor between words, then select Split. If there is no cursor position, the app chooses a word break near the playhead. Merge next joins the next block in time order. Every edit supports undo.
- Apply block rules reorganises your current text using the character limit and sentence setting. Original words retain their estimated speech timestamps. Text that you edited or imported uses timing estimated within its existing block. You can undo the result.
- Download the final SRT for editing or translation elsewhere. Save a project to keep editable timings and word data. A project does not contain your media; reopen the original file when you resume.
Keyboard shortcuts outside text fields: Space to play or pause, ← / → to seek one second, Shift plus an arrow to seek five seconds. Use Ctrl / Cmd + Z for undo and Ctrl / Cmd + Shift + Z for redo outside text fields.
The file extension alone does not determine support. MP4 and MOV audio is extracted locally. Video playback still depends on the codecs supported by your browser. If a video cannot be displayed, you can still transcribe its supported audio and use audio playback. Very large files can exceed browser memory. Use shorter clips or a separate WAV file on memory-limited devices. If using separate audio, it must start at the same point as the video.
Only app files and speech models are requested from this website. No media or text is sent to a server. There are no ads, analytics, accounts or third-party requests. Your browser may cache model files. Only your model and block preferences are saved in browser storage; the transcript is not automatically saved. Use Save project before closing.