Lossless Screen Recording, Right in Your Browser

Strictly, lossless means zero data loss between source and recorded file. In a browser environment that's impossible end-to-end — the final MP4 has to use a widely-supported codec, and those codecs are lossy by design. So when we say lossless, we mean what matters in practice:

• No re-encoding during recording. The audio path captures raw PCM and writes it straight to OPFS. The video path uses the browser's hardware encoder once, with avc3 in-band parameter sets so resolution shifts mid-recording don't produce a broken file.
• No double-compression. Most cloud-backed recorders compress on-device, then re-compress on their server. Two lossy passes compound. RecATab does one encode, at the end, when FFmpeg WASM muxes the final file.
• No third-party processing. Your recording never touches a server unless you click Share. Everything — capture, mixing, muxing — runs in your browser's sandbox.

The screen capture comes from getDisplayMedia — the standard browser API for tab and window capture. From there:

• Video. Goes through the browser's hardware H.264 encoder (5 Mbps target, avc3 codec descriptor) and writes fragmented MP4 chunks to OPFS.
• Audio. Bypasses the encoder entirely — an AudioWorklet captures 32-bit float PCM samples at the device's native rate and streams them to disk untouched.
• Optional microphone. Input flows through the same AudioWorklet graph and is mixed with tab audio at PCM precision.
• When recording stops. FFmpeg WASM muxes the video (copy, no re-encode) and the PCM audio (encoded once to AAC at high quality) into the final MP4.

That's one encode pass for video, one encode pass for audio. No intermediate lossy compression, no server round-trip, no quality drift.

If you record talking-head demos at moderate volume, most recorders are fine. Lossless capture pays off when:

• You're recording music, podcasts, or anything with dynamic range
• You'll re-edit the audio later (compression artifacts compound when re-encoded)
• Sibilance, breath, or room tone matter for tone of voice
• You're sharing recordings to people who themselves edit them