← back to features
AUDIO ap-audio sox + ffmpeg

Podcast & Audio
Post-Production

Upload recording → AI denoise & edit → Export broadcast MP3

① Upload raw WAV/MP3 ② Agent denoises & normalizes ③ Broadcast-ready MP3

// AGENT SESSION LOG

U

"Clean up episode 12 of my podcast. Remove background noise, normalize loudness to -14 LUFS, trim silence longer than 2 seconds, and export as 128kbps MP3."

AP

// Analyzing audio: 47m 23s, 96kHz stereo WAV

  • → Detect noise profile from silence sections
  • → Apply noise reduction (sox noisered)
  • → EBU R128 loudness normalization to -14 LUFS
  • → Trim silence segments > 2s
  • → Convert to 128kbps MP3 via ffmpeg
agentputer — ap-audio executor

$ ap-audio analyze --input ep12-raw.wav

✓ Duration: 47m 23s · Peak: -2.1 dBFS · Noise floor: -45 dB

$ ap-audio denoise --input ep12-raw.wav --profile auto

✓ Noise profile sampled · Applied: -18dB noise reduction

$ ap-audio loudnorm --target -14 --tp -1.5 --lra 11

✓ Integrated loudness: -23.4 LUFS → -14.0 LUFS (EBU R128)

$ ap-audio trim --remove-silence --threshold 0.02 --min-duration 2.0

✓ Removed 4m 12s of silence · Final: 43m 11s

$ ap-audio convert --mp3 128k --output podcast-ep12-final.mp3

✓ Exported: podcast-ep12-final.mp3 · 41.7 MB · 128kbps

elapsed: 2m 18s
🎵 podcast-ep12-final.mp3 128kbps · 43m 11s

// WHAT THIS SCENARIO SUPPORTS

🔇

Adaptive denoising

Auto-samples noise profile from silent sections. Works for HVAC hum, room tone, and recording hiss.

📊

EBU R128 loudness

Industry-standard loudness normalization for podcast platforms (Spotify -14 LUFS, Apple -16 LUFS).

✂️

Silence detection & trim

Configurable threshold and minimum duration. Remove dead air without touching speech.

🔄

Format conversion

WAV, FLAC, OGG, AAC, MP3 — any codec, any bitrate. Batch convert a whole folder.

🎚️

Multi-track mixing

Mix intro/outro music with voice track. Set independent volume levels per track.

📋

Audio analysis report

Get duration, peak levels, loudness, and clipping stats before processing.

// DEPLOY YOUR AGENT

Professional audio. Zero manual editing.

Upload the raw recording. Get broadcast-ready audio. No DAW, no plugins, no hourly rate for an audio engineer.