# Transcribe (https://openclip.app/docs/tools/transcribe/)



The Transcribe tool converts speech into accurate text. One run gives you all three output formats at once: a plain transcript, an SRT subtitle file, and a VTT subtitle file.

Open it at [https://openclip.app/tools/transcribe](https://openclip.app/tools/transcribe), or press Cmd-K and type "transcribe".

## What it accepts [#what-it-accepts]

Transcribe takes both video and audio files. It is the only tool family that accepts audio input directly, so podcast episodes and voice memos work as well as videos.

Input limits follow your plan: 200 MB and 10 minutes on free, 2 GB and 6 hours on paid. See [Tools overview](https://openclip.app/docs/tools/overview/) for the full quota table.

## Output formats [#output-formats]

| Format | File    | Use it for                                    |
| ------ | ------- | --------------------------------------------- |
| Text   | `.json` | Reading, editing, or feeding into other tools |
| SRT    | `.srt`  | Subtitles for most video editors and players  |
| VTT    | `.vtt`  | Subtitles for web players and HTML5 video     |

You do not pick a format up front. Every run produces all three, and you download whichever you need from the result screen.

## Options [#options]

| Option          | Type   | Default | What it does                                                                            |
| --------------- | ------ | ------- | --------------------------------------------------------------------------------------- |
| Detect speakers | On/off | On      | Labels who is speaking (diarization), so multi-person recordings read as a conversation |

Detect speakers sits under the advanced options. Leave it on for interviews and podcasts. Turning it off can make sense for single-speaker recordings where labels add noise.

## Running a transcription [#running-a-transcription]

1. Open the tool page.
2. Upload a file, or pick a video already in your Library.
3. Check the Detect speakers setting if you need to change it.
4. Run. The page polls the job until it completes.
5. Download the text, SRT, or VTT file.

## Watermarks and quota [#watermarks-and-quota]

Transcripts are text, not visuals, so they are never watermarked, even on the free plan. Each run still counts against your daily tool quota (10 runs on free, 500 on paid).

## Related [#related]

* Want captions burned into a video instead of a subtitle file? That is the clipping pipeline, not this tool. Start at [What is OpenClip?](https://openclip.app/docs/).
* Need to pull the audio out of a video first? Use [Convert](https://openclip.app/docs/tools/convert/) with the Extract audio operation.