Frame-level reconstruction · no blur, no crop
Remove text from videowithout ruining the shot.
Captions, titles, price tags, date stamps — if it is burned into the picture, we take it out and rebuild what was behind it. The frame keeps its size, the audio stays untouched.
First run free · no card · files deleted after 7 days
What this application does
VideoTextRemove is a web application for removing text that has been rendered into video frames — burned-in captions, subtitles, titles, watermarks, date stamps and channel handles. You upload a clip you own, mark the text or let detection find it, and download a new MP4 in which the background behind the letters has been reconstructed. Nothing is blurred, cropped or covered with a sticker.
A Google account is used only to sign you in: it creates your account, holds your credits and shows your name in the dashboard. Reading the site never requires one — pricing, the published test results and every guide page are open to anyone.
More detail: About · Acceptable use · Privacy policy
What it is not
- Not a generator. It reconstructs background pixels inside the area you mark and cannot add people, objects or scenes that were not already in the file.
- Not a face or likeness tool. There is no face swap, no undressing, no voice synthesis. Creating or altering non-consensual intimate imagery is prohibited and ends the account.
- Not for other people’s work. Removing a watermark or credit from footage you do not own is a copyright problem, not a feature.
Try it on your own clip
Upload, mark the text, download. Processing runs server-side — you can close the tab.
MP4 · MOV · WebM · MKV — up to 100 MB per file
Background rebuilt, not blurred
The model generates replacement pixels using motion information from neighbouring frames. Flat walls and simple textures reconstruct almost invisibly.
Framing stays exactly the same
No crop, no scale, no letterbox. The frame keeps its resolution, so the shot still cuts with the rest of your edit.
Automatic or manual
Auto detection handles centred captions in the lower half of the frame. For corner stamps, logos and styled text, drag a box and restrict the repair.
Including the caseswhere it does not work.
We start from footage with no text in it, burn the captions on ourselves, and send it to the model with no manual hints. Because we know what the clean frames look like, we can measure how much of the picture comes back — and show you the failures.
See the before and after, with test conditions
Three steps, no timeline
Upload a clip
MP4, MOV, WebM or MKV, up to 100 MB. Nothing is installed.
Mark the text
Let the model find captions, or draw a box over the text you want gone.
Download clean
Review in motion, then export at the original resolution. Close the tab if you like — the job keeps running.
Find your case
Every page below covers a different kind of burned-in text, with the specifics that change how the removal behaves.
Remove text from video, without blurring the frame
Subtitle removerRemove subtitles from video
Hardcoded subtitle removalRemove hardcoded subtitles from video
Watermark removalRemove a text watermark from your own footage
Timestamp removalRemove a date stamp from old video
TikTok text removalRemove text from a TikTok video you own
GuideHow to remove text from video: three methods, compared
Reels text removalRemove text from Instagram Reels
YouTube text removalRemove text from a YouTube video
Logo removerRemove a logo from video
No blur, no mosaicRemove text from video without blurring it
iPhone workflowRemove text from a video on iPhone
No-crop subtitle removalRemove subtitles without cropping the frame
Dual-language subtitle removalRemove bilingual subtitles from video
Complex background removalRemoving subtitles over a moving background
Questions people actually ask
How do I remove text from a video?+
Upload the clip, choose automatic detection or drag a box over the text, then submit. Processing happens on a server and usually takes a few minutes. When it finishes you download a file at the same resolution you uploaded.
Is it free?+
Every new account gets 1 free credit on signup — enough for a one-minute video, no card required. After that, processing costs from $0.08 per minute of video, sold as a subscription or a one-time pack that never expires.
Can it remove text without blurring?+
Yes. The area is reconstructed rather than covered. Results depend on the background: simple surfaces look untouched, while faces, foliage and fast motion may show artefacts you should check before publishing.
What video formats are supported?+
MP4, MOV, WebM and MKV, up to 100 MB per file. Length is capped per job: 60 seconds on the free credit, 60 minutes on any paid plan — a longer clip has to be split, and credits are charged per minute either way. The output keeps the frame size of your source; frame rate is standardised to 30 fps.
Do you keep my videos?+
Your uploaded file is deleted 7 days after you send it, on every plan; the processed result lasts 7 days on Free and Starter, 30 days on Plus, 90 days on Premium. You can delete a finished job from your dashboard and both files are removed immediately. We never use your footage to train models, and nobody here watches it. The privacy policy names the exact deletion rules, so you can check the claim instead of taking our word for it.
What do I get back?+
An MP4 at the resolution you uploaded, with the audio track copied through untouched. Frame rate is standardised to 30 fps, so a 24 fps source comes back at 30 — worth knowing if the clip has to cut against 24 fps footage.