A long graphite transport deck between two spools carries a strip of dark tiles under a reading head; an ejector has lifted three tiles into a side tray, and past a diagonal cut the take-up spool and the rest of the deck continue as a ghosted outline, with one sky-blue line running the length of the rail.
Other

AudioEditor

Cleans spoken audio via Whisper transcription, Claude segment classification, and ffmpeg cuts, plus breath taming and noise-gate artifact repair

01
Workflow
09
Tools
01
Reference
20
Triggers

The Problem

Raw recordings are full of garbage: filler words, dead air, false starts, stutters. You can manually edit in a DAW, but that takes hours per episode and you still end up making judgment calls on every pause — is that silence rhetorical or accidental? Generic AI transcription tools will give you a transcript but won't touch the audio. And most automated cutters are blunt: they'll slice out a deliberate pause for effect the same way they cut a dead-air gap where you lost your train of thought.

How This Skill Approaches It

Run the full pipeline: Whisper transcribes at word-level timestamps, then Claude classifies each segment as KEEP, CUT_FILLER, CUT_FALSE_START, CUT_STUTTER, or CUT_DEAD_AIR — distinguishing a rhetorical pause from accidental silence before any cut is made. ffmpeg executes the edits with 40ms qsin crossfades at every cut point and fills gaps with extracted room tone so the edits don't click. Breaths get attenuated to 50% rather than removed, because fully cutting them sounds unnatural. An optional Cleanvoice API pass handles mouth sounds and loudness normalization. Three modes let you tune aggressiveness: --preview shows proposed edits before touching the file, --aggressive tightens thresholds for denser cleanup, --polish adds the Cleanvoice final pass.

Not for video composition (HyperFrames) or tight-cutting a podcast or spoken recording for publishing (the video skill's TightCut intent (maintainer-side; not in the public release))

In Action

What you say to your DA, and what the AudioEditor skill actually does.

  • You say "clean up the audio on this podcast recording"
    Runs the Clean workflow: Transcribe.ts → Analyze.ts (Claude classifies every segment) → Edit.ts (ffmpeg cuts with 40ms crossfades and room-tone fill) → outputs cleaned MP3.
  • You say "show me what edits you'd make before you touch the file"
    Runs Clean with --preview: transcribes and classifies all segments, shows the proposed cut list with timestamps and reasons, makes no changes until you approve.

Inside the Skill

The thinking, frameworks, and architecture that distinguish this skill from a generic version of the same task.

Ideal State

  • Every output keeps what the speaker meant, their deliberate pauses and natural breathing included, and no edit can be heard.
  • Done means heard: a clean scan or a zero exit code never counts on its own.
  • Repairs are minimal-touch: each tool changes only the defect it targets and leaves the rest of the audio alone.
  • Nothing is changed or uploaded without the user's say: previews modify nothing, ambiguous file matches are asked about, and cloud polish runs only on request.

Intent Routing

Intent Trigger File
Clean "clean audio", "edit audio", "remove filler words", "clean podcast", "remove ums", "cut dead air", "polish audio" Intents/Clean.md

Standalone breath, mouth-click, gate-tick and loudness requests ("ticking audio", "gate artifacts", "loud breaths", "mouth clicks", "tame the breaths") also route to Clean, using the Tools table and Gate Artifacts below.

File Read when
Tools/ each tool's <Name>.help.md holds its full flags and exit codes
Setup.md first-time setup: ffmpeg and Whisper to install, the optional Cleanvoice key, and a synthesized-tone proof

Data & Config

User customizations live at <configRoot>/LIFEOS/USER/CUSTOMIZATIONS/SKILLS/AudioEditor/ (configRoot is $CLAUDE_CONFIG_DIR when set, else ~/.claude). When that directory exists, its PREFERENCES.md, configurations and resources apply and override the defaults here; when it is absent, the defaults stand. First-time setup (accounts, keys, config, a proof it works) is in Setup.md.

Tools

Tool Command Purpose
Transcribe bun ${LIFEOS_SKILL_DIR}/Tools/Transcribe.ts <file> Word-level transcription via Whisper
Analyze bun ${LIFEOS_SKILL_DIR}/Tools/Analyze.ts <transcript.json> LLM-powered edit classification
Edit bun ${LIFEOS_SKILL_DIR}/Tools/Edit.ts <file> <edits.json> Execute cuts with crossfades + room tone
Polish bun ${LIFEOS_SKILL_DIR}/Tools/Polish.ts <file> Cleanvoice API cloud polish
Pipeline bun ${LIFEOS_SKILL_DIR}/Tools/Pipeline.ts <file> [--polish] Full end-to-end pipeline
GateScan bun ${LIFEOS_SKILL_DIR}/Tools/GateScan.ts <file> [--json] Detect noise-gate ticking (silence-boundary steps); exit 1 on defects
GateRepair bun ${LIFEOS_SKILL_DIR}/Tools/GateRepair.ts <in> <out.mp4> --finalize [--abr 192k] Repair gate ticking; --finalize iterates until the encoded file scans clean
BreathTame bun ${LIFEOS_SKILL_DIR}/Tools/BreathTame.ts <file> [--dry-run] [--json] Attenuate individual breath/mouth-click events in word gaps; word-confined, ducks never removes, proves bit-identity outside events
LoudnessLock bun ${LIFEOS_SKILL_DIR}/Tools/LoudnessLock.ts <in> [--out <out.mp4>] Measure or lock delivery loudness to −14 LUFS / −1dBTP (YouTube standard); self re-measures, exit 0 only in tolerance

Gate Artifacts

Gate checks run only when someone actually reports ticking or clicking; routine per-export scans were retired 2026-07-15. On a report: GateScan the file to confirm (sample-domain steps, exit 1 on defects), then GateRepair --finalize to fix. Repair before leveling when possible, and scan the final encode, not the intermediate WAV. If a RAW recording scans dirty, a capture-chain gate is back on; surface it.

API Keys Required

Service Env Var Where to Get
Cleanvoice (for polish step, optional) CLEANVOICE_API_KEY cleanvoice.ai Dashboard Settings API Key

Gotchas

  • Cloud polish uploads audio to external service. Confirm the user is okay with cloud processing for sensitive content.
  • Capture-chain noise gates cause the ticks. Recorder filters and macOS Voice Isolation truncate audio to digital zero with no fade, and leveling amplifies each edge into an audible tick. The class is fixed at capture: remove any noise-gate filter from the recorder's mic chain. (2026-07-13: hundreds of gate edges in one recording reached published audio.)
  • AAC re-introduces steps near silence. An intermediate WAV that scans clean says nothing about the encoded file, which is why GateScan runs on the final encode. (2026-07-13)
  • dB-domain cliff detectors false-positive on fade feet. A legitimate cosine fade into digital zero has infinite dB slope at its foot, so any >NdB/2ms detector flags it forever. Verify repairs with sample-domain STEP detection (GateScan), never dB slopes. (2026-07-13)
  • One-sided iterative fades don't converge on stray boundary impulses. A single sample sitting on a gate boundary survives repeated one-sided fades; fade-down-then-up "mutes" preserve a blip's edges and its leading step. The V-notch (cosine to zero at the boundary, both sides) and hard-zeroing inside gated silence are the converging fixes. (2026-07-13)
  • Meter-clean ≠ ear-clean. Scanners find ticks; repairs can create audible holes the tick-scanner calls clean (the 2026-07-12 duck incident). Match the probe to the defect class a human hears, keep repairs minimal-touch, and give the principal before/after listen clips at swap time.

Examples

  • "clean up the audio on this podcast file" → Clean, full pipeline (transcribe → analyze → edit) → cleaned MP3 with filler words, stutters, and dead air removed.
  • "show me what edits you'd make to this recording" → Clean with --preview → proposed edit list, audio untouched; the user reviews, then runs again to apply.
  • "aggressively clean this audio and polish it" → Clean with --aggressive --polish → tighter filler thresholds, then the Cleanvoice pass for mouth sounds and normalization.

Workflows & Routing · 1

Each workflow is one job the skill runs. The trigger phrases route your request to the right one — this is the skill's routing table.

  1. 01
    Clean Intents/Clean.md

    clean audio, edit audio, remove filler words, clean podcast, remove ums, cut dead air, polish audio

Tools · 9

Deterministic executables the workflows call — the code that does the real work, not prompt scaffolding.

  • Analyze.ts
  • BreathTame.ts
  • Edit.ts
  • GateRepair.ts
  • GateScan.ts
  • LoudnessLock.ts
  • Pipeline.ts
  • Polish.ts
  • Transcribe.ts

How to Invoke

Say any of these to your DA and LifeOS activates the AudioEditor skill automatically:

  • "clean audio"
  • "edit audio"
  • "remove filler words"
  • "clean podcast"
  • "remove ums"
  • "cut dead air"
  • "polish audio"
  • "trim recording"
  • "cut stutters"
  • "ticking audio"
  • "clicking audio"
  • "audio clicks"
  • "gate artifacts"
  • "popping audio"

Or invoke explicitly:

Skill("AudioEditor")

References · 1

Auxiliary files the skill loads at runtime — frameworks, guides, configs.

  • Setup

Want LifeOS to do this for you?

Install LifeOS on your machine — your DA gets the AudioEditor skill plus 96 others, all hooked into one Life OS.