How the Podcast Editor Works
AnyTool mixes podcasts entirely in the browser with two engines. Each file is decoded once with the Web Audio API into a cached AudioBuffer and a peak waveform lane; a live preview schedules each track via AudioBufferSourceNode.start(when, offset) through per-track gain, EQ and compression, with solo/mute and an AnalyserNode-driven auto-duck that lowers music under the voice bus. Remove-silence detects low-RMS spans and rejoins the speech locally. The final mix renders in an OfflineAudioContext, then a self-hosted single-threaded ffmpeg.wasm engine (bundled core, no CDN) applies EBU R128 loudnorm (-16 podcast / -14 Spotify), per-track noise reduction (highpass + afftdn) and encodes MP3 (libmp3lame), M4A (AAC), OGG (libvorbis), FLAC or WAV; a WAV-only export skips ffmpeg for an instant result, and a target-MB mode plans the bitrate.
- Runs client-side — all tracks are never uploaded
- Sync-accurate multi-track timeline with per-track volume/solo/mute/trim
- One-click remove-silence and EBU R128 normalize (-16/-14 LUFS)
- Per-track noise reduction (afftdn) and auto-ducking
- Self-hosted ffmpeg core (bundled) — no CDN
- Export MP3/M4A/OGG/FLAC/WAV with a target-size option
