Everything it does, and why

Recording

Capture a browser tab, an application window or a whole screen, with your camera in a corner bubble and your microphone mixed in. Tab audio comes through too, so a demo with sound stays a demo with sound.

It uses the browser's own screen-capture API — nothing to install and nothing running in the background when you aren't recording. Chrome, Edge, Firefox and Safari 17+ on a desktop are all supported. A three, two, one counts you in, so the take does not start with you reaching back to the mouse.

Upload that finishes when you do

Most tools begin uploading at the moment you press stop, which is why a long recording ends with a progress bar. Zoomin buffers the stream into 5 MB parts and pushes each one to storage as it fills.

By the time you stop, only the last few seconds are left to send. If your tab crashes mid-recording, the parts already sent are still there — and a sweep releases the abandoned upload rather than leaving you billed for it.

Playback that works on a phone

Browsers record VP8 or VP9 in a WebM container. Safari and iOS handle that badly or not at all — and a phone is where a link you texted someone actually gets opened.

So every recording is transcoded to H.264 in an MP4, with the index moved to the front of the file. That last part is why it starts playing immediately instead of downloading the whole thing first.

Cards rendered into the video

Nine intro animations and nine matching sign-offs are in your account from the start, so the first branded recording does not begin with an empty page and a design job. Upload your own as well — stills held for a few seconds or short clips, with up to four text layers each: eight typefaces, nine-point placement, timing, and fade, rise or slide entrances.

The card holds the look; the recording holds the words. Type what this one should say, set how long it runs, and the same sting opens an entire series while saying something different each time. A preview plays the card with your text on it at the exact size, position and timing the encoder will use — so you are not guessing and then waiting for a transcode to find out.

They are encoded into the video itself in the same pass as the transcode. Nothing is overlaid by a player, so they survive being downloaded, embedded or forwarded.

A camera bubble you can put anywhere

Your camera sits in a corner bubble, and on Pro you can drop an image behind it. Seventeen come with the account — offices, meeting rooms and abstract backdrops — or use your own. It is composited as you record, not bolted on afterwards, so there is no green screen to set up and no second application running.

Because it goes into the video itself, it survives everything the video survives: the download, the embed, the copy somebody forwards.

Sharing you can take back

Videos are private by default and stored in a private bucket; every URL the app hands out is signed and expires. Publishing is an explicit per-video decision — having a link is not enough on its own.

Changed your mind? Rotating the link kills every copy already sent, which is the only real unshare once a URL is loose. On Pro you can add a password, an expiry date, or a readable name — /v/spring-launch rather than a line of hex — which is the difference between a link you can say out loud and one you have to paste.

Enough analytics, not more

You see how many people watched and when, plus a retention curve — the share of viewers still watching at each tenth of the video. If they leave in the first ten seconds, that is worth knowing, and it is on the free plan rather than behind the upgrade.

Repeat loads from the same visitor within an hour count once, so the figure means something. We do not fingerprint viewers, build profiles, or tell you who watched — a share link is not a tracking pixel. The curve tells you about the video, not about the person.