Screen Recorder

Screen recording used to mean installing software. Browsers now expose a capture API that asks your permission, shows you exactly what will be shared, and hands the page a live video stream — which is enough to build a recorder that needs no installation and no account.

The practical consequence is that the recording never leaves your machine. A screen recording is one of the most sensitive files a person can make — it can contain messages, passwords being typed, client work, anything that was on screen — and uploading it to a stranger's server to be processed is a strange thing to be asked to do.

This tool captures, buffers the data in the page, and saves it. There is no upload step to opt out of, because there is no upload.

The capture API is deliberately constrained. A page can ask, but the browser draws the picker, and the user chooses what is shared. A page cannot pre-select your screen, cannot start capture without a click, and cannot keep capturing silently — every browser shows a persistent indicator for the entire session.

Audio is the messiest part, and the differences are the browser's, not this tool's. Sharing a browser tab generally offers that tab's audio. Sharing an entire screen offers system audio on Windows in Chromium browsers, and usually does not on macOS, where the operating system has no route for it without a virtual audio device. Firefox does not offer system audio at all. If sound matters, share a tab.

Microphone audio is separate and always available, subject to its own permission prompt. When both exist they have to be combined before recording, because a recorder writes one audio track — so the page mixes them through Web Audio first.

The output container depends on what the browser can write. Where MP4 is available it is used, because a recording is usually something you send onwards and MP4 opens everywhere. WebM is the fallback, and is equally valid — it simply has narrower support in older editing software.