AI video transcription
Transcribe video to text you can actually use.
Generate a timestamped transcript from a YouTube link or uploaded video, translate it when needed, and export it as TXT, SRT, VTT, or CSV.
Review it. Translate it. Subtitle it. Build with it.
Transcript
- 00:04The thing people miss about transcription
- 00:07is that the text is the starting point, not the deliverable.
- 00:11Once you have the words with timing attached,
- 00:14you can search them, translate them,
- 00:17turn them into subtitles, or cut the video with them.
- 00:21That is the part that actually saves you a week.
Low-res previewExport
The useful part of a video should not be trapped in playback.
Finding one quote, reviewing an interview, creating subtitles, or searching a two-hour recording should not require repeatedly scrubbing through a timeline.
Video only
- Watch
- Pause
- Rewind
- Type it out
- Search manually
- Reformat for subtitles
Then all of it again for the next quote, the next clip, the next language.
With reap
- Add video
- Generate transcript
- Review with timestamps
- Export the format you need
Video and text, together
Read the transcript without losing the moment.
Review generated text beside the source video and use timestamps to connect every line with the moment it was spoken.
Transcript beside the video
The generated text sits next to the source, so you are never reading words with no idea where they came from.
Timed text
Every line carries the moment it was spoken, which is what makes a transcript navigable instead of just long.
Easier quote verification
Before you publish a quote, jump to its timecode in the source and hear it said.
Timestamps on or off
Keep them while you work, hide them when you want a clean block of prose to read or paste.
The demo at the top of this page is the same idea: press play, watch the spoken line highlight, and toggle timestamps off when you just want to read. Step-by-step guide in the help center.
One transcript. The right format for the next job.
The same generated text leaves reap in whichever shape the next tool expects. Pick a format to see how the output differs.
Clean text for reading, notes, writing, and repurposing.
Plain prose with no timing markup. The format to paste into a doc, a draft, or a research file.
The thing people miss about transcription is that the text is the starting point, not the deliverable. Once you have the words, you can search them, translate them, and cut the video with them.
Multilingual transcripts
Transcribe the language spoken. Read it in the script you need.
Choose the source language, optionally translate the transcript, and select native or romanized script for supported languages, so Hindi, Urdu, Arabic, and Bengali speech can be read as Hinglish, Romanized Urdu, Arabizi, or Romanized Bengali instead of a script your audience may not read. Accuracy varies with the language, the accent, and the recording. Clear audio and the correct source language give the strongest result.
Native script
The language in its standard writing system.
मैंने वीडियो का ट्रांसक्रिप्ट बना लिया है।
Roman script
The spoken language represented with Latin characters.
Maine video ka transcript bana liya hai.
Both lines above are the same Hindi sentence. Roman script changes how the words are written, not what language they are. It is not an English translation. Translation is a separate, optional setting, and Roman output is available on supported languages. The current list of all 100+ is in the supported-language reference.
Video to transcript in four steps.
Add your video
Paste a YouTube link you are authorized to use, or upload a supported video file.
Choose the language
Select the spoken language, add a translation target if you need one, and pick native or Roman script.
Generate and review
Read the transcript beside the video, with timestamps on when you need to locate a moment.
Export
Download TXT, SRT, VTT, or CSV for the next part of your workflow.
The start of something useful
A transcript should lead somewhere.
Use the transcript as the foundation for clips, captions, translated subtitles, editing, dubbing, and content automation. Each stage is a step you configure and start, so nothing downstream fires on its own.
- Source video
- Transcription
- AI dubbing
- AI clipping
- Captions
- Editing
- Export or publish
Transcription at scale
Transcribe from code, or from your agent.
Everything the transcript workflow does is also a REST API, and the same actions reach Claude, Cursor, and any other MCP client through reap’s MCP server. API, CLI, and MCP draw the same plan credits as the dashboard, from $9.99/mo.
- Create a transcription in one call
- Point at an uploaded file or a source URL, name the spoken language, and reap returns a project you can follow.
- Translate and pick the script
- Set a translation target and native or Roman output on the same request, instead of running a second pass.
- Track every project
- Check a project for processing, completed, or failed, and pull the finished text as TXT, SRT, VTT, or CSV.
- Get told, instead of asking
- Register a webhook and reap posts to your endpoint the moment a transcription reaches a final state, with no polling loop to run.
Each key runs at 10 requests per minute, and credits are shared 1:1 with the dashboard. Every endpoint and field is in the create-transcription reference and the webhook guide. reap produces the transcript data that search, indexing, and analysis systems run on. It is the text layer for those systems, not a search engine in itself.
Claude
Cursor
ChatGPT
VS Code
OpenClaw
Who it’s for
Useful anywhere spoken video becomes information.
The pattern is the same everywhere: a recording that holds something worth reading, and no time to watch it back to find it.
Creators
Turn a video into captions, an article draft, show notes, and social copy.
Podcasters
Review a long interview and find the quote you half-remember.
Educators
Turn a recorded class into readable lesson material and subtitle files.
Marketing teams
Repurpose webinars, customer interviews, demos, and events into text.
Researchers
Work through recorded material with timestamps you can cite.
Developers
Turn a video collection into structured data your systems can index.
Agencies
Process client footage at volume and hand back the format each expects.
Journalists
Search a two-hour recording instead of scrubbing it a second time.
Accessibility teams
Produce timed text as the foundation for more accessible video.
Course creators
Give every module a transcript students can read, search, and skim.
Event teams
Turn a day of talks into an archive somebody can actually read.
Localization teams
Start from timed source text instead of transcribing it by hand.
Frequently asked questions
AI video transcription converts the speech in a video into written text automatically. reap generates a timestamped transcript from a YouTube link or an uploaded video, lets you review it beside the source video, optionally translate it, and export it as TXT, SRT, VTT, or CSV so the text can be read, subtitled, searched, or used in the next step of a workflow.
Yes. Paste a YouTube link you are authorized to use and reap transcribes the audio, or upload the video file directly. Only transcribe content you own or have permission to use.
You can paste a YouTube link or upload a supported video file. Accepted file types, maximum file size, and maximum duration depend on your plan, so check the transcript guide in the reap help center for the current limits rather than assuming.
reap supports 100+ languages for transcription and the same coverage for translation, including Arabic, Bengali, Chinese, French, German, Hindi, Indonesian, Japanese, Korean, Portuguese, Spanish, Swahili, Tamil, Telugu, Thai, Turkish, Urdu, and Vietnamese. Accuracy is not identical across every language, accent, recording condition, or subject area, so review the output before you publish it. The full list is in the reap help center.
Yes. Choose the language spoken in the video, then optionally pick a translation target. The transcript is generated in the translated language, and the same TXT, SRT, VTT, and CSV exports are available.
Roman script writes the spoken language using Latin characters instead of its standard writing system, so Hindi becomes Hinglish, Urdu becomes Roman Urdu, and Arabic becomes Arabizi. It is not a translation into English: the words are the same, only the script changes. Select Native for the original writing system or Roman for Latin characters on supported languages.
Yes. The transcript review view has a timestamps toggle. Turn them on to connect each line to the moment it was spoken, or off for a cleaner block of text to read and copy.
TXT for clean reading text, SRT for timed subtitles, VTT for web captions, and CSV for structured rows you can analyze. You can also download the extracted source audio where supported.
Yes. Export SRT or VTT and upload the file wherever your video lives, whether that is a platform, a player, or an editor. If you want captions burned into the video with styling and animation instead of a sidecar file, that is reap's AI Captions feature.
In reap's video editor, transcript text is an editing surface: you can set a clip start or end from a word, correct a word, highlight it, add an emoji, remove a caption, remove a section of video, or cut. That is a capability of the editor rather than of the standalone transcript export screen.
Yes. POST to https://public.reap.video/api/v1/automation/create-transcription with an uploadId or sourceUrl, an optional language, an optional translationLanguage, and a transcriptionScript of native or roman. The project returns a status you can poll, output includes word-level timing, and webhooks fire when a project reaches a final state.
Yes. AI handles the first pass well when the speech is clear and the source language is set correctly, but names, numbers, acronyms, product terms, technical vocabulary, and passages with background noise or overlapping speech are worth checking before you publish or quote the transcript.
Your video already contains the words. Put them to work.
Generate a timestamped, translatable transcript and export it in the format your next workflow needs. Plans on the pricing page.
Transcribe a videoExplore reap