# Claudacity

> Claudacity

- Skill: `kimiguel/claudacity` (Agent Skill, multi-file: 7 files)
- Install (CLI): `npx skillmds@latest add kimiguel/claudacity`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kimiguel/claudacity/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: KiMiGuel (https://skillmd.com/u/kimiguel)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kimiguel/claudacity

---


# Claudacity

AI mixing and mastering guide for Audacity. Controls Audacity via mod-script-pipe and teaches technique calibrated to the user's declared genre.

---

## Genre Setup (required on first use)

Before any playbook runs, ask:

> "What genre are you mixing/mastering? (e.g. grindcore, hip-hop, jazz, metal, folk, punk, electronic...)"

Save the answer as `GENRE` for the session. All mixing/mastering advice adapts to that genre — dynamics expectations, EQ targets, loudness norms, and processing philosophy. If the user doesn't know, ask what the closest reference artist sounds like instead.

---

## Session Menu

On `/claudacity` or trigger phrase, show:

```
Claudacity — Audacity AI Session
Genre: [GENRE or "not set — type your genre to begin"]

1. Normalize / Gain Stage
2. Mix
3. Master

Type a number or describe what you need.
```

Load the matching playbook. Apply genre context throughout.

---

## Startup Checklist

Before sending any pipe commands:

1. **Audacity open?** — If `GENRE` is set and user is ready, check pipe. If Audacity not open: `open -a Audacity` and wait.
2. **Pipe present?** — Run `scripts/pipe_test.py`. If pipe absent: tell user to enable it in Audacity > Preferences > Modules > mod-script-pipe, then restart Audacity.
3. **Round-trip verify** — Send `GetInfo: Type=Tracks` and confirm response before any processing commands.

---

## Pipe Protocol

Named pipes at:
- `/tmp/audacity_script_pipe.to.501` — write commands here
- `/tmp/audacity_script_pipe.from.501` — read responses here

Command format: one command per write, terminated with `\n`. Read until blank line.

Common commands:
```
GetInfo: Type=Tracks
SelectAll:
Normalize: PeakAmplitude=-3
Amplify: Ratio=<float>
ExportAudio: Filename=/path/to/output.wav Format=WAV
Macro: <MacroName>
```

---

## Audacity Macros

Macros are plain text files in `~/Library/Application Support/audacity/Macros/`.
Triggered via pipe: `Macro: <MacroName>` (filename without extension).

To create a macro for a genre-specific processing chain: write the macro file, then trigger it via pipe.

---

## Loudness Analysis

Run `scripts/loudness_check.py <file>` before any processing. Outputs:
- Peak dBFS
- RMS dBFS
- Crest factor (peak - RMS in dB)

Use these numbers as the baseline. Run again after mastering to verify targets hit.

Crest factor reference by genre: high dynamic range (jazz, folk, classical) = 18–25dB; rock/punk/metal = 11–16dB; heavily limited pop/EDM = 6–10dB. Ask the user if unsure what's appropriate for their genre.

---

## Loudness Targets by Distribution

| Format | Integrated LUFS | Peak ceiling |
|---|---|---|
| Bandcamp / CD | -14 to -7 LUFS (genre-dependent) | -0.1 dBFS |
| Streaming (Spotify/Apple) | -14 LUFS | -1.0 dBFS |

Ask the user where they're releasing before finalizing master loudness. Aggressive genres (metal, punk, hardcore) typically target -9 to -7 LUFS on Bandcamp. Dynamic genres (jazz, classical) typically land -16 to -14 LUFS even on Bandcamp.

---

## WAV Fingerprinting

Run `scripts/wav_chunks.py <file>` to identify RIFF chunks before mastering. Flags:
- **Already mastered** (bext/umid/DGDA chunks): warn before reprocessing
- **Logic bounce** (cue/ResU/LIST chunks): confirmed mix-down, safe to master

---

## Teaching Philosophy

Explain the **why** behind every move. Don't just say "boost 2kHz" — say what it does to the sound and why it matters for the user's genre. One action at a time. Confirm results before moving to the next step.

---

## Hardware Context

If the user mentions their monitoring setup, factor it into advice. Common issues: near-fields against reflective walls (add low-end buildup), consumer earbuds (rolled-off highs), laptop speakers (no sub information). Ask if unclear.