How the Voice Recorder Works
AnyTool records the microphone entirely in the browser. Capture uses getUserMedia and the MediaRecorder API (with a device picker via enumerateDevices and echo-cancellation/noise-suppression/gain options), while a Web Audio AnalyserNode draws a live waveform and level meter; there is no recording time limit. On stop the take is decoded to an AudioBuffer for a post-record waveform with draggable trim handles and one-tap silence auto-trim. WAV is written directly in JavaScript and WebM is the raw recording, both instant; MP3 (libmp3lame), M4A (AAC, +faststart), OGG (libvorbis) and FLAC are encoded by a self-hosted single-threaded ffmpeg.wasm engine (bundled core, no CDN), with optional loudness normalize.
- Runs client-side — the recording is never uploaded
- Self-hosted ffmpeg core (bundled) — no CDN
- Live waveform + level meter; no recording time limit
- Device picker and noise-suppression/echo-cancellation toggles
- Post-record waveform trim + silence auto-trim
- WAV/WebM instant, plus MP3/M4A/OGG/FLAC export
