Upload an MP4, MOV, MKV or audio file and get the full transcript in minutes. OpenAI Whisper transcribes the audio itself, so recordings that were never captioned work exactly the same as ones that were. Download timestamped SRT or VTT subtitles, or copy the plain text. 3 free lectures a day, no account, nothing to install.
How It Works
No account to create and nothing to install. Drop the file in exactly as it came off the camera, phone, Zoom recording or screen recorder.
Drag an MP4, MOV, MKV, WebM or AVI onto the upload box, or pick an MP3, WAV, M4A, FLAC or OGG if you only have the audio. Up to 500 MB and 3 hours long. No format conversion first — the file goes in as it is.
OpenAI Whisper listens to the audio rather than reading an existing subtitle track. It handles accents, background noise and technical vocabulary, and auto-detects the language across 100+ of them.
The full transcript appears in the Transcript tab with a search box, so any phrase is one query away. Copy it as plain text in one click, or download it as a timestamped SRT or VTT file.
Formats
If your file is not on this list, it is not silently rejected after a long upload — the format is checked before anything is sent.
MP4, MOV, MKV, WebM, AVI. The audio track is what matters, so screen recordings, phone footage, Zoom and Teams exports and camera files all behave the same.
MP3, WAV, M4A, FLAC, OGG. Useful when you recorded a talk on a phone or a dictaphone and never had video at all.
SRT for adding subtitles to a video — the most widely supported subtitle format there is. VTT for web players, which expect it. Both carry a timestamp on every caption. For the words alone, copy the plain text straight from the Transcript tab.
Why This One
Most of them cap you at 10 minutes
A lecture, a deposition, a two-hour interview or a recorded meeting is exactly the thing you cannot afford to chop into six-minute pieces and stitch back together. Files here run up to 3 hours, and a long one counts the same as a short one — one lecture out of your 3 a day, whatever its length.
Auto-captions read subtitles; this reads audio
Tools that pull an existing caption track produce nothing at all on a file that was never captioned, which is every file that came off a camera or a screen recorder. Whisper transcribes the sound, so there is nothing to be missing.
Find the moment, not just the words
A transcript you cannot navigate is only slightly better than the video. The SRT and VTT downloads carry a timestamp on every caption, and the notes built from the same upload mark each key point and chapter with the minute it came from, so you can go straight to that point in the video.
You get more than the transcript
The same upload also produces chapter-by-chapter notes, timestamped key points, key terms with definitions, a generated quiz, and an audio recap you can listen to instead of rereading. The transcript is the floor here, not the ceiling.
A failed job costs you nothing
If a transcript fails to generate it is not counted against your 3 a day. Your daily allowance is automatically restored.
FAQ
Everything you need to know about turning a video file into text.
What video files can I upload?
MP4, MOV, MKV, WebM and AVI for video, and MP3, WAV, M4A, FLAC and OGG for audio. The limit is 500 MB per file and up to 3 hours of runtime on the free tier. Nothing needs converting first — upload the file exactly as it came off your camera, phone, Zoom recording or screen recorder.
Do I have to sign up or install anything?
No. Uploading a file and downloading the transcript works in the browser with no account and nothing installed. You get 3 free lectures a day. A free account only adds a library that keeps your transcripts across devices.
Does it work on videos with no captions or subtitles?
Yes, and that is the point of it. OpenAI Whisper transcribes the audio itself rather than reading an existing subtitle track, so a raw camera file, a Zoom recording or a screen capture that has never been captioned transcribes exactly the same as one that has.
How accurate is it, honestly?
Whisper reaches roughly 95% word accuracy on clear speech. It degrades on three things: several people talking over each other, heavy background noise, and unfamiliar proper nouns or jargon. Anything shown on screen but never spoken will not appear at all — it listens rather than reads. The full transcript is always shown so you can check anything that looks wrong.
Which export format should I pick?
SRT if you are adding subtitles to a video — it is the most widely supported subtitle format there is. VTT if the video will play on the web, since HTML5 players expect it. Both are downloads with a timestamp on every caption. If you only need the words, for notes or to feed another tool, copy the plain text from the Transcript tab instead — that copy has no timestamps.
Can I use a YouTube link instead of a file?
Not directly, and we would rather say so than waste your time. YouTube blocks downloads originating from datacentre servers, which is what every web app runs on, so our server cannot fetch the audio. Install our free Chrome extension, which reads the caption track from your own browser session, or download the video and upload the file here. There is a fuller explanation on the YouTube transcript page. Loom, Dailymotion and direct MP4 or MP3 links all work without any of this.
Is my video kept on your servers?
Files are deleted within 24 hours of processing, and nothing you submit is used to train AI models. See the Privacy Policy.
What happens if a transcript fails?
It is not counted against your 3 a day. Your daily allowance is automatically restored — you are never charged for something that did not work.
3 free lectures a day, no sign-up, and failed jobs are not counted automatically.
Upload a Video Free →