Guide · Agent skill
How to use the AI Audiobook Narration agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The ai-audiobook-narration skill is a reviewed catalog entry that lets an agent such as Claude Code, Codex or OpenClaw turn a finished manuscript or course text into chapter audio, starting with one sample chapter. The synthesis runs on the paid, hosted Beatra service, not on your computer, so this guide explains what the skill does, how to install it safely, how to ask for it in plain language and what to check before any money is spent. Read it through once before installing, because the package it points to runs scripts and makes network calls.
AI Audiobook Narrationsickn33/agentic-awesome-skills · MIT Download the skill Quick start
- 1Confirm the fit. Check that you have finished text, accept that it will be sent to a third-party paid service, and that you do not need local TTS, M4B files, translation or mastering.
- 2Download and unzip the skill. Get the 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.
- 3Review the pinned package. Read every file in the package, including scripts and hooks, and confirm the network and credential behaviour before you let the agent run anything.
- 4Test the connection. Run the free verify check to confirm the scripts can reach the service without spending credits.
- 5Authorize with Beatra. Approve the browser sign-in as a separate decision, knowing it stores a bearer token locally.
- 6Request a sample chapter. Ask the agent to narrate one sample chapter, then review the cost card and approve or decline that single paid call.
- 7Review and continue. Listen to the sample, check the reported charged credits, and only then approve any further chapters or refinements.
What the skill makes the agent do
Agent skills are folders of instructions and optional scripts that an AI agent loads when a task matches them. This one covers audiobook and course narration: it sends a finished text to the Beatra service, produces ordered chapter audio with one narrator, and then supports focused refinements.
The catalog entry itself contains no executable code and performs no Beatra operation. The actual package it points to bundles three standard-library Python scripts that make network calls and store a sign-in credential.
- Turns a final manuscript or course text into chapter-by-chapter audio
- Produces a representative sample chapter first so you can review the narrator
- Does not do local or offline text-to-speech, M4B packaging, translation or mastering
Install the skill from the download page
Download the skill ZIP from this page and unzip it. For Claude Code, place the unzipped folder in your user skills directory at ~/.claude/skills/ or in a project's .claude/skills/ folder. If you use the Claude apps instead, upload the skill under Settings, then Capabilities, then Skills.
The installation steps in the package fetch its pinned source with Git from github.com, check the expected tree ID, and then copy the reviewed folder into the skills directory. Ask for approval before each of the two stages, because the second one changes agent configuration. Start a new agent session if the host only discovers skills at startup.
- Git is required to fetch the pinned package from github.com
- Python 3.10 or newer is required, using the standard library only
- Stop if the expected tree check does not print the confirmation line
Review the package before you activate it
Before activation, read every file in the pinned package, including each script, reference and hook, and check its network and credential behaviour. The package lists these files for review: SKILL.md, manifest.json, install-source.json, the bundled Markdown references, and the scripts authorize.py, mcp_client.py and uninstall.py.
Automated scans are only inventory aids. Do not treat them as proof that the code is safe. If you cannot complete a full review, stop before copying or running anything.
- Confirm the package has no symlinks, executable bits or binaries
- Check the self-updater entry points are absent, as the package describes
- Do not rely on a digest or a grep result as evidence of safety
git -C src ls-tree -r 4f5f6b5ec10ee989be010c4990a7a087cc3c721c -- skills/ai-audiobook-narration | awk '$1 != "100644"'Connect your Beatra account
Authorization is a separate decision. It opens a browser sign-in with your Beatra account and stores a bearer token on your machine. Only run it after you have agreed to use the paid service.
You can test the connection without spending credits. The verify command is described as a free, non-billable check.
- You need a Beatra account and paid credits for synthesis
- The token is stored locally, so uninstall it when you stop using the skill
python3 "$HOME/.claude/skills/ai-audiobook-narration/scripts/mcp_client.py" verifyTrigger the skill with natural requests
You do not need special syntax. Say what you want narrated, that it should be paid hosted work on Beatra, and that you accept sending the text to a third party. The skill is meant for finished text only, so draft manuscripts are a poor fit.
Expect the agent to read the current model information and show a cost card for exactly one paid call before it uploads anything. Planning or saying just do it does not count as approval, so answer the cost card explicitly.
- Narrate chapter one of my finished manuscript on Beatra, starting with a sample chapter, and show me the cost first
- Make an audiobook from this final course text, one narrator, and wait for my approval before each paid step
- Check my Beatra connection without using any credits
Understand the cost and approval loop
The workflow is a loop. The agent reads free discovery information, shows the cost card for a single call, uploads only the files you supplied for that task, submits once with a stable request identifier, and then reports the delivered audio and the net charged credits.
If a response is lost, the package says to retry only with the same request identifier and identical arguments. Do not submit a replacement while a task is still running. Ask again before any further paid stage.
- Cost card shows the tool, model, estimate and files to be uploaded
- Results report billing.net_charged_credits so you can track spend
- Voice cloning should only be used with a sample whose speaker has documented consent
Update, uninstall and remove the skill
This pinned version has no self-updater and does not download or replace its own files. To move to a newer version, repeat the review steps for that version, compare the file changes against your installed copy and replace the folder only after approval. A version not listed in the catalog entry has not been reviewed.
To uninstall, run the dry-run first, then the real command, and read the JSON decision it prints. The output says whether this was the last Beatra skill on the device and whether the token was revoked.
- Dry run first, so you can see what will be removed
- A disconnected result means the token was revoked when the revoked flag is true
python3 "$HOME/.claude/skills/ai-audiobook-narration/scripts/uninstall.py" --dry-runWhat people use it for
- Narrating a finished novel or memoir chapter by chapter with one consistent narrator
- Producing a sample chapter to judge a narrator before committing to a full book
- Turning a finished online course script into audio lessons
- Refining a single chapter after review without re-running the whole book
Frequently asked questions
How do I install the ai audiobook narration skill?
Download the skill ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. In the Claude apps, you can upload it under Settings, then Capabilities, then Skills. Start a new session if the host only loads skills at startup.
Is the ai audiobook narration skill free?
The skill itself is listed under the MIT licence in the catalog, and the package's SKILL.md names MIT-0 (MIT No Attribution). Synthesis is not free: it uses paid Beatra credits, and the skill shows a cost estimate before each paid call.
What do I need to run it?
You need Git to fetch the pinned package from github.com, Python 3.10 or newer using only the standard library, a Beatra account with browser sign-in, and paid credits for synthesis.
Does it work offline or on my own computer?
No. The narration is produced on the hosted Beatra cloud service, and the skill does not do local or offline text-to-speech. Your text and any voice sample are sent to that third-party service.
Is it safe to install?
The catalog entry contains no executable code, but the package it points to runs Python scripts that make network calls and store a sign-in token. Read every file before activation and only proceed if you accept those network and credential behaviours.
How does it compare with local text-to-speech tools?
Local tools keep your text on your machine and usually avoid per-use charges, but this skill is designed for paid, hosted narration. It does not do M4B packaging, translation or mastering, so choose a local tool if you need those or cannot send manuscripts to an outside service.
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
Use this skill if you have a finished text, want paid hosted narration, accept sending it to a third-party service, and have reviewed the pinned package's scripts and credential storage. Pick a local text-to-speech tool or another option if you need offline work, cannot share manuscripts or voice samples externally, or want to avoid per-use charges.
Download the skill