Upload a clip, mark the burned-in text, and download a clean MP4 with the background rebuilt. Audio, duration, and frame size stay unchanged.
PRE-RENDERED SAMPLE
THIS CHANGES EVERYTHING
or drop a file here
No clip handy? Try one of these
This demo renders in your browser. Your file never leaves your device.
Before / after
The cleaned version plays first. Hover, press, or use the corner button to see the original.
Dialogue
One caption line sits over the iron bars. The rebuild keeps the rail edges straight instead of smearing them into the sky.
5s · Auto · Tears of Steel (CC BY 3.0)
Dialogue
White text on a dark shoulder is the forgiving case: the fill has almost no detail to invent.
5s · Auto · Tears of Steel (CC BY 3.0)
Bright pass
A flare burns right behind the words. Check the halo at both ends of the cue before publishing.
5s · Auto · Tears of Steel (CC BY 3.0)
Moving shot
Both cues are replaced in one pass while the camera and the subject keep moving.
5s · Auto · Tears of Steel (CC BY 3.0)
Watermark
The logo and @handle shift position while a burned caption changes. Both are gone and the crowd behind them is rebuilt.
5.3s · Auto · TikTok @nasa
Watermark
The watermark sits on crater texture and title lettering. The repair keeps the ground detail instead of leaving a soft patch.
3.9s · Auto · TikTok @nasa
Plays the cleaned clip. Hover, press, or tap the corner button to see the original.
Built for burned-in text
Captions, handles, logos, and timestamps cannot be switched off after export. CleanVideoText tracks them across frames and rebuilds the background behind every word.
Hardcoded subtitles
Remove dialogue captions baked into interviews, courses, and reposted clips — not just subtitle tracks.
Learn moreTikTok & Reels
Clean downloaded TikToks and Reels without re-editing or covering the frame with a bar.
Learn moreUsernames & handles
Erase handles, chat text, and follower counters that sit on the picture.
Learn moreLogos & product text
Remove a logo or on-pack text when you have the rights to the footage.
Learn moreTimestamps & bugs
Clean dates, channel bugs, and camcorder timestamps from archive footage.
Learn moreThe toolkit
Pick the kind of text you are dealing with. Every mode runs on the same clip, the same audio, and the same frame size.
Subtitles
Social clips
Branding
ArchiveHow it works
No timeline, no keyframes, no export settings. The workflow is built around one job: remove what is burned in.
Drag a file in or choose it from your device. We check the format, size, and duration before anything is charged.
Use Auto to scan the whole frame, or Select area to repaint a specific box. Everything inside a box is affected, not just text.
We rebuild the background, keep the audio, and re-encode an MP4 with the same frame size. Review it before you publish.
30 free credits every month. No card required.
Free every month
The sample clip is free in your browser. Every account gets 30 free credits each month: 30 seconds in Auto, or 15 seconds in Select area. It is a metered free tier, not unlimited.
What it handles
Five kinds of text, one workflow. If it is visible in the frame and you have the rights to edit it, start with a box or subtitles mode.
The sample never uploads. Live jobs go to the processing provider; see the Privacy Policy for retention.
Dialogue baked into the picture, including translated captions. Soft subtitle tracks are a different problem — remove those with a remux, no AI needed.
Captions and stickers saved into a TikTok, Reel, or Short. Works best when the text stays in one area for several seconds.
Creator tags, LIVE badges, chat snippets, and follower counters. Moving handles need a tight box or the subtitles workflow.
Brand marks and on-pack copy on footage you own or are licensed to use. We do not accept third-party watermark removal.
Camcorder dates, channel marks, and old tape stamps. Low-contrast text on detailed backgrounds is the hardest case.
CASES & LIMITS
The sample video above is pre-rendered. These scenarios explain which control fits the text and what to inspect before a live render.
EASIER
01Lower-band captions over a steady wall or soft-focus scene give the model more nearby detail to rebuild.
MOVING TEXT
02Auto scans the frame for text. Use Select area when you only want to target a fixed region.
OWNED FOOTAGE
03A tight Select area box can target a fixed username or sticker, but everything inside the box is repainted.
HARD CASE
04There may be too little nearby information to rebuild the covered area accurately.
The video above is a pre-rendered browser sample, not output from the live engine and not a render of your uploaded clip.
Compare the routes
Four ways to deal with burned-in text, and where each one actually works.
Swipe to see all columns
Comparison describes workflows, not measured output quality. Re-editing from the original project is still the cleanest option when you have it.
We do not cover text with a shape. The area is rebuilt and the rest of the frame stays untouched.
The original audio track and timing are preserved in the export.
The output keeps the source aspect ratio, so it drops back into your edit.
Format, size, duration, credits, and processing time are shown before you start.
Resources
Guides, comparisons, and workflows for the text you cannot toggle off.
Not sure which route fits your file? Start with how to remove text from a video, compare blur-free methods, or go straight to removing hardcoded subtitles.
Learn
Practical routes for subtitles, handles, and archive footage — including when not to use AI.
SubtitlesTell soft subtitles from burned-in captions, then pick the cleanest route for your file.
MethodsCrop, bar, blur, or rebuild: what each method does to the shot, and when it is the right call.
WorkflowFrom re-exporting the source project to rebuilding the pixels, with limits for each.
FAQ
Straight answers about modes, limits, audio, privacy, and what happens when the background is too complex.
Still unsure? Email us before you spend credits.
support@cleanvideotext.comUpload the clip, choose Auto to scan the full frame or Select area to draw boxes, then download the clean MP4. The live engine processes the whole uploaded clip, and credits follow its length.
The sample clip is free in your browser. Every signed-in account gets 30 free credits each month: 30 seconds in Auto, or 15 seconds in Select area. The uploaded clip must fit the quota, and each job has a 5-second minimum. Exports keep the source resolution and carry no CleanVideoText watermark.
Yes. The sample demo and the free monthly credits use the same workflow: the original audio and duration are preserved and the output keeps the source frame size and resolution. Check the file before publishing.
Yes. A worker rebuilds pixels behind text across the clip. Auto scans the full frame for text-like regions; Select area sends a mask and repaints everything inside your boxes, not just text.
That is the main use case. CleanVideoText rebuilds the background instead of covering it, so there is no blur bar, mosaic, or crop. On complex backgrounds the rebuild can still leave artifacts, so review the result before publishing.
Processing time depends on clip length, resolution, and background complexity. Keep the tab open until the download is ready; there is no email notification or background queue yet. The credit cost is shown before you start.
Auto scans the full frame for text-like regions. Select area repaints everything inside the boxes you draw, so keep boxes tight and away from details you want to preserve.
Sometimes. Short, slow movement with a clean background works best. Fast movement, text over faces, or text next to a logo you need to keep will need a tighter box and a review — and the result may still need manual cleanup.
The workflow is designed to preserve the original audio and duration, then re-encode the MP4. Always check the downloaded file before publishing.
MP4 or MOV, up to 200 MB and 10 minutes. Trim larger or longer files first. The live engine processes the whole uploaded clip, so its length is what gets charged.
Not in the live engine yet. It processes the whole file you upload, and the credit total follows that clip length. Trim a five-second test clip or the final segment in your editor before uploading.
Not inside the app. A soft subtitle track is a separate stream, so remove it with a remux in FFmpeg, HandBrake, or your editor — no AI is needed. The AI workflow is for text burned into the picture.
Auto uses 1 credit per second; Select area uses 2 credits per second because it sends a mask. Each job has a 5-second minimum. Every account gets 30 free credits each month, and packs are listed on the pricing page.
The sample never uploads. Live jobs are sent to the processing provider only to run the requested job; see the Privacy Policy for retention. We do not use your footage for model training, and we do not send media, file names, or mask coordinates to analytics.
No. We only support footage you own or are licensed to edit. Removing a third-party watermark to redistribute someone else’s work is not allowed.
Built for burned-in text
Sign in with Google, upload a clip, and see what the background looks like after the text is gone.
CASES & LIMITS
The sample video above is pre-rendered. These scenarios explain which control fits the text and what to inspect before a live render.
Fixed dialogue subtitles
Review the first and last frame of each subtitle.
Word-by-word social captions
Fast motion and busy scenes can soften the repaired patch.
Handles and stickers
Use footage you own or are licensed to edit.
Ask AI
A ready-made question about burned-in subtitles, soft tracks, and the trade-offs opens in a new chat.
The question is prefilled where the AI supports it. If not, it is copied to your clipboard — paste and send.
