Turn video files into clean Markdown.
Recorded meetings, lectures and demos become structured documents: summary, timestamped transcript, and notes on what appears on screen.
Your files are never stored: documents convert in your browser where possible, and anything sent to our AI travels encrypted (HTTPS) and is discarded after processing. Privacy Policy
A video is the least searchable format you can store knowledge in. CleanMD converts MP4, MOV, WEBM and MKV files (up to 100 MB) into structured Markdown: a summary, a transcript with [mm:ss] timestamps split into chapters, speaker labels — and because the AI also watches the video, short *[On screen: ...]* notes when slides, demos or whiteboards matter for understanding. Your video is streamed for processing and never stored.
Working with related formats? See also Audio to Markdown, YouTube to Markdown.
When the screen matters as much as the audio
In a screen recording, a lecture with slides or a product demo, half of the information is on screen and never said out loud. A transcript of the audio alone loses the code that was typed, the numbers on the slide and the button that was clicked.
CleanMD sends the video to an AI model that watches and listens. The document you get has a summary, the cleaned transcript with a [mm:ss] timestamp at the start of each paragraph, and notes marked as [On screen: ...] where something visible matters, such as a slide title, a chart or a line of code. Topics become headings, so a long recording can be scanned instead of replayed.
Uploads are streamed straight to the model instead of being stored on our servers first. That's what makes longer files possible: a 19-minute video takes around ten minutes to process, and the connection stays open the whole time. There's a hard stop at 15 minutes of processing, with a clear error instead of a page that waits forever, and Cancel stops the work on the server too.
The summary is in the language spoken in the video. If the answer comes back cut off at the model's length limit, the request is retried automatically, so a long recording doesn't end in the middle of a sentence.
MP4, MOV, WEBM and MKV files up to 100 MB are accepted. Video is part of Premium, sharing the daily allowance with audio and YouTube: 10 conversions and 120 minutes of media. If the video is on YouTube already, paste the link instead, since the model can then read it directly without an upload.
How it works
- 1Drop your video file (MP4, MOV, WEBM, MKV — up to 100 MB) in the converter above.
- 2The AI watches and listens, then structures everything: summary, timestamped transcript, visual notes.
- 3Preview the Markdown and download the .md file.
Before and after
[22-minute screen recording] One presenter, slides + live demo. No captions, no transcript.
## Summary Sprint demo of the new export flow, with a live walkthrough. ## Transcript **[00:15]** Today I'll show the new export pipeline. *[On screen: architecture diagram]* **[02:40]** The queue handles retries automatically...
Frequently asked questions
Which video formats are supported?
MP4, MOV, WEBM and MKV, up to 100 MB per file.
What are the visual notes?
When something shown on screen matters — a slide, a chart, a live demo — the transcript gets a short italic note like *[On screen: architecture diagram]* at the right spot.
Is my video uploaded and stored?
It's streamed for processing and deleted immediately after the conversion. Nothing is persisted — recorded meetings deserve that guarantee.
Is video to Markdown free?
No. Video is processed by AI and costs by the minute, so it's part of Premium: files up to 100 MB, up to 10 audio, video or YouTube conversions and 120 minutes of media a day. Converting documents and images stays free.