Guide · Agent skill
How to use the Audacity audio editing CLI agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
This guide explains how to install and use the cli-anything-audacity agent skill, which lets an AI agent edit audio in Audacity through a text command-line interface instead of a graphical window. It is written for readers who have never used agent skills. By the end you will know what the skill does, how to install it, how to ask an agent for edits, and what safety points to check first.
Audacity Audio Editing CLIHKUDS/CLI-Anything · Apache-2.0 Download the skill Quick start
- 1Check prerequisites. Confirm you have Python 3.10 or newer and Audacity installed on your computer.
- 2Install the package. Run pip install cli-anything-audacity in your terminal.
- 3Test the CLI. Run cli-anything-audacity --help and create a test project with project new to confirm it works.
- 4Download and place the skill. Download the skill ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 5Make a request. Ask the agent in plain language, naming the audio task and the full file paths.
- 6Review and render. Check the commands the agent runs, then ask it to render and confirm the output file exists.
What does the Audacity CLI skill let an agent do?
The skill teaches an AI agent to drive Audacity, a free audio editor, through a command-line interface. The CLI is described as a stateful tool for audio editing, following the same patterns as the GIMP and Blender CLIs in the same repo.
Once installed, the agent can create project files, import and add audio clips, trim and split clips, add and configure tracks and effects, set labels and selections, undo or redo steps, and render the project to an audio file. Output can be returned as JSON so the agent can read results reliably.
- Project: create, open, save, inspect and print project JSON
- Track: add, remove, list and set name, mute, solo, volume or pan
- Clip: import, add, remove, trim, split, move and list clips
- Effect: list available effects, add, remove, set parameters and list
- Selection, label and media commands for ranges, markers and file checks
- Export: list presets, show preset details and render the project
- Session: status, undo, redo and history
Install the skill in Claude
Download the skill ZIP from this page. You can then install it in one of two ways. For Claude Code, unzip it into your personal skills folder ~/.claude/skills/, or into a project's .claude/skills/ folder if you want the skill limited to one project. In the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
The skill's own instructions say the CLI itself is installed as a Python package. Install it with the command below. You need Python 3.10 or newer, and Audacity must already be installed on your system.
- Requirements: Python 3.10 or newer
- Requirement: Audacity installed on the system
- Package: cli-anything-audacity
pip install cli-anything-audacityCheck the CLI works before asking an agent
Run the help command to confirm the CLI is on your path. Then create a test project. The skill's examples use -o to set the output file name. Using --json makes the output machine-readable, which is the format the agent uses.
Start with a new project so you can see the basic flow before running edits on real recordings.
cli-anything-audacity --help
cli-anything-audacity project new -o project.json
cli-anything-audacity --json project info -p project.jsonHow do I trigger the skill with natural requests?
You do not need to type commands yourself. Ask the agent in plain language and name the audio task and the files involved. Use absolute paths for files, because the skill's guidance for agents asks for them. The agent should then choose the right commands from the groups listed above.
Good requests say what the result should be. For example: create a project, add a clip, trim the start, add a fade-style effect, and render to a file. Be clear about where the output should go and whether you want the original file left untouched.
- "Create a new Audacity project at /home/me/podcast/intro.json and import /home/me/podcast/intro.wav onto track one."
- "Trim the first 2 seconds from the clip in that project and show me the project info."
- "List the available effects, then add one to track 1 and tell me the effect index."
- "Undo the last change and show me the session history."
- "Render the project to /home/me/podcast/intro_final.wav and confirm the file exists."
How the CLI works under the hood
The CLI keeps session state, so each command builds on the last. Undo and redo cover up to 50 levels of history. Projects are saved and loaded as JSON, and the CLI tracks changes made in a session.
You can also start the interactive REPL by running the command with no subcommand. In that mode you type commands one at a time, with tab-completion and history, and use undo and redo to move through steps. The REPL is useful for checking what the agent did.
- Undo and redo: up to 50 levels
- Project files: saved and loaded as JSON
- Output modes: human-readable by default, or --json for agents
cli-anything-audacity
cli-anything-audacity --project myproject.json export render output.wavWhat safety points should I check first?
The skill runs scripts, and the agent may run commands on your computer. The tool writes and renders files locally. It does not use the network, according to the skill's listing. Because of this, avoid giving the agent sensitive recordings on shared machines, and check which folders it writes into.
For agent use, the skill asks the agent to check return codes (0 means success, non-zero means error), read stderr for error messages on failure, and verify that output files exist after an export. Review the commands an agent runs before you approve them, especially any that write or overwrite files.
- Keep backups of original recordings before editing
- Use absolute paths so you know exactly which file is changed
- Confirm outputs exist after rendering
- Be careful with overwrite options when exporting
What people use it for
- Batch-trimming the start and end of many recordings into a consistent length
- Building a repeatable podcast intro or outro project that an agent can rebuild on demand
- Adding labels or markers at set times in a long recording for later review
- Scripting an audio processing pipeline that reads JSON results from each step
- Rendering a finished project to a file as the final step of a content workflow
Frequently asked questions
Is the Audacity CLI agent skill free?
The skill's listing is published under the Apache-2.0 licence, and Audacity itself is described on this page as a free audio editor. Check the licence file in the repository for the exact terms before reusing the code.
What do I need to install the Audacity CLI skill?
You need Python 3.10 or newer, Audacity installed on your system, and the pip package cli-anything-audacity. The skill depends on both Python and Audacity, so it will not work without them.
How do I install the Audacity agent skill in Claude?
Download the ZIP from this page. For Claude Code, unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. In the Claude apps, upload it under Settings → Capabilities → Skills.
Does the skill need an internet connection?
The listing says the skill does not use the network. It runs scripts and writes and renders files locally on your computer.
Can I use Audacity CLI without a graphical window?
Yes, that is the point of the skill. It is a text-command interface with no waveform view, so it suits scripted and repeatable edits. If you want to edit by sight and hand, use the Audacity application directly.
Why does my agent need --json output?
The skill's guidance for agents says to always use the --json flag, because it returns structured output the agent can parse. Human-readable output is the default and is harder for an agent to read reliably.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits creators and developers who want an AI agent to make scripted, repeatable edits to audio projects, and who are comfortable installing a Python package and using a terminal. It is not a good choice if you want a visual, hands-on editor, if you cannot install Python packages or Audacity, or if you need to process sensitive recordings on a shared machine, since the agent may run commands on your computer.
Download the skill