Qué se publicó, y cuándo.

Remove Green Screen now has a dedicated page with the same upload and recent-clips workspace as Remove Video Background.

A new Remove Green Screen page helps you process imperfect green-screen footage with AI, without manual chroma key controls. It runs the same processing as Remove Video Background and shares the same free allowance, so a folded cloth, an uneven shade or a little spill does not send you into a keying session.

Remove in the top menu now opens onto both: Remove Video Background for footage shot anywhere, and Remove Green Screen for footage shot against green.

The agency example now shows one consistent presenter across both finished versions.

Colour and Blur controls now wait for your choice before updating, with clearer progress and retry feedback while the preview is being prepared.

The German, Spanish, French and Portuguese pages still described paid plans, ten-minute clips, batch uploads and API access. None of that exists. Those pages now state the same limits as the English ones: three free background removals per 24-hour window for each signed-in account, clips up to 60 seconds, transparent WebM at the size you uploaded, files kept for 24 hours, and background changes that are free and unlimited. The menu item that read Pricing in those four languages now reads Free Access, and the pages it leads to no longer list features that were never built.

Two English pages were behind as well. The guide to removing a video background said the background could only be chosen later in another editor, and the product video page answered a flat “yes” on reflective and glazed items. Both now match what the tool actually does and what it actually struggles with — clear glass and transparent packaging included.

On the home page, the workspace used to show “3/3 left in this window”, which never said left of what. It now reads “3/3 free removals remaining in this 24-hour window”, with the real number in every state, and it appears once instead of alongside a second sentence saying the same thing.

The four translated home pages also link to their own language’s use case pages instead of sending you to the English ones, and their walkthroughs now mention that you can put a colour, blur, image or video behind a finished clip here and export an MP4, free, without using one of your three daily removals.

The four translated home pages gained the illustrated sections the English one already had. There are now three walkthroughs — one transparent master across several finished scenes, one cutout across every aspect ratio, and a cleaner handoff from whoever films to whoever edits — each with real stills rather than diagrams.

Three questions were answered alongside them, in the language you are reading: what an AI video background remover actually is, which footage works best, and why a transparent file can still look like it has its original background when you open it directly.

Adding a new background now has its own page at /change-background/, open to anyone rather than tucked inside the workspace.

It opens the way the home page does: the tool on the left, your finished clips on the right, so a video you have already made transparent is one click from a new background. Underneath, the page shows the same subject on four different backgrounds, walks through the three steps from original footage to exported MP4, explains which background suits which job, and answers the questions people ask most.

Two things it now states plainly. The image or video you choose as a background is combined with your clip inside your browser and is never uploaded. And changing a background does not use another video allowance — the allowance is spent when the original background is removed, so you can keep trying different scenes on the same clip.

The three illustrated explanations on the homepage now use direct, everyday wording instead of editing jargon. They walk through how to reuse one transparent video across different backgrounds, how to fit the same recording into wide, square and vertical frames, and how a recording is handed from whoever films it to whoever finishes it.

The homepage now shows the workflow instead of only describing it. Three illustrated sections follow one cut-out subject through the places she ends up: composited over a photographed room, a product set and a presentation screen; placed into widescreen, square and vertical frames with room left for headlines or captions; and handed from the original clip through an editor to the finished exports on a laptop and a phone.

The stills are demonstration composites rather than customer footage, and the page says so. The person in them is generated, and every scene was assembled from a single cut-out placed over separate backdrops, which is why she is identical in all of them.

The questions people arrive with have moved into the FAQ, where they are easier to scan: what an AI video background remover is, which footage gives the best result, and why a transparent WebM can still look like it has its original background when you open it straight from your desktop. Formats, limits and replacing a background are each answered in one place now instead of two.

The homepage now answers the question we were asked most often: why a transparent WebM can still look like it has its original background when you open it straight from your desktop. That is usually the player not applying the alpha channel rather than a failed cutout, so there is now a section on how to check transparency properly, and what to do when an editor will not read the file.

Compatibility is described more carefully everywhere. Instead of saying the output works in a list of named apps, pages now state what the file actually is — WebM with a VP9 alpha channel — and that whether transparency displays depends on the browser, or on the editor’s version and platform. Where a limitation is documented by the vendor, it is linked.

The homepage also explains what an AI video background remover is, which footage gives the best result, what a transparent clip is useful for, and the formats and limits, all on the page itself.

The use-cases overview is now a page you can choose from rather than three links: each workflow says who it is for, what goes in, what the transparent result is used for afterwards, and what to watch for in your footage, with a short guide to picking between them.

Breadcrumbs on the tool pages no longer include a middle step that had no page behind it.

Explanations across the site now lead with what you see rather than how it is built, from the transparency guide on the tool pages to the steps on the homepage.

Your account and workspace both show the exact date and time your allowance resets, in your own time zone, instead of a date alone.

Signed-in accounts can complete up to 3 successful video background removals per 24-hour allowance window. Each video can be up to 60 seconds long. Output preserves the source dimensions and is delivered as watermark-free transparent WebM using VP9 with alpha. Source files and generated results are retained for 24 hours.

Entries below this one describe earlier configurations. Where they state a resolution tier, a seconds budget, or an output format, that detail has been superseded by the terms above.

Historical entry. See the current free access entry for what applies today.

The maximum length of a free clip was raised, because the previous cap was short enough to refuse most of the footage people actually wanted to process.

Historical entry. See the current free access entry for what applies today.

The daily free allowance was increased, because one clip a day was enough to see whether the matte works but not enough to find where it breaks.

Historical entry. See the current free access entry for what applies today.

Output format became a choice made at export rather than at upload, so switching format costs nothing in quality and does not use the allowance again.

Historical entry. See the current free access entry for what applies today.

First public release: upload a video, get back a file with a real alpha channel, with automatic subject detection and Google sign-in.