More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    acx audiobook audio checker

    1

    Check every audiobook chapter against ACX's audio requirements before you upload: RMS, peak, noise floor, room tone, 192 kbps CBR and more.

    $7

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    acx audiobook audio checker

    acx audiobook audio checker

    Example session with this skill installed

    I'm about to upload my audiobook to ACX. My edited WAV chapters and a couple of MP3 test exports are in the chapters folder. Will they pass? Tell me what to fix.

    • Read your context and instructions
    • Compiled the acx audiobook audio
    • Generated the data export

    2 of 8 files pass. Fix these before uploading:

    FileProblemFix
    ch02_too_quiet.wavRMS -26.6 dB (needs -23 to -18)Limiter at -3.5 dB, then raise gain into the window
    ch03_noisy_short_head.wavNoise floor -52.0 dB (max -60); head room tone 0.15 s (needs 0.5-1 s)Find the noise source or apply gentle noise reduction to room tone; paste 0.5 s of room tone at the start
    ch04_hot_peaks.wavPeak -0.5 dB (max -3); tail 6.55 s (max 5 s)Limit to -3.5 dB; trim the tail
    ch05_stereo_48k.wav48,000 Hz (needs 44.1 kHz); 0.80 s of digital silenceRe-export at 44.1 kHz; replace silence with room tone
    ch02_vbr.mp3VBR (128-256 kbps frames)Re-export at 192 kbps CBR
    ch03_128k.mp3128 kbps CBR (min 192)Re-export at 192 kbps or higher

    Project-wide: mono and stereo files are mixed (ACX requires all mono or all stereo), and RMS varies 8.4 dB between chapters. Even them out.

    Passing: ch01_pass.wav (RMS -21.1, peak -4.0, noise -68.1, head 0.80 s, tail 2.55 s) and ch01_192k_cbr.mp3 (192 kbps CBR, MPEG-1 Layer III, 44.1 kHz).

    Full details are in acx-report.md and acx-report.csv. Re-run after fixing, then confirm in ACX Audio Lab.

    (Numbers come from a real run on the bundled test files.)

    acx-report.csv

    CSV · data export

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Catch ACX rejections before you upload, on every chapter at once.

    acx audiobook audio checker gives your AI agent a pre-flight check for audiobook files headed to ACX (Audible). Point it at a folder of WAV masters or MP3 exports. It measures each file against ACX's published audio submission requirements and returns a pass/fail table, the exact number behind every failure, and the usual fix. It never modifies your audio.

    What it checks

    • Levels (WAV): RMS between -23 and -18 dB, peak no higher than -3 dB (with a warning above -3.5 dB, because MP3 encoding can push peaks up), noise floor at -60 dB or lower, and clipped samples.
    • Spacing (WAV): 0.5 to 1 second of room tone at the head and 1 to 5 seconds at the tail, plus generated digital silence where clean room tone should be.
    • Format (WAV and MP3): 44.1 kHz sample rate, length up to 120 minutes per file.
    • MP3 exports: reads every frame header to confirm MPEG-1 Layer III, 192 kbps or higher, constant bitrate (catches VBR exports frame by frame), channel mode, duration and the 170 MB file limit. It also flags a LAME Xing (VBR) tag on a file that should be CBR.
    • Whole project: all mono or all stereo, and the RMS spread between chapters so loudness stays consistent.
    • Reports: terminal table, acx-report.md and acx-report.csv. Exit code 1 when anything fails, so it also works as a gate before upload.

    Problems it solves

    • Files bounced by ACX QA days after upload
    • Not knowing which of 40 chapters has the noise or level problem
    • A master that peaks at exactly -3.0 dB and then fails after MP3 export
    • Accidental VBR or 48 kHz exports from a DAW preset
    • Head and tail room tone that's too short, too long, or pure digital silence

    Who it's for

    Self-published authors narrating their own audiobooks, indie narrators and producers on ACX, small audiobook studios, and KDP authors expanding into audio.

    What's included

    • SKILL.md: how to gather files, run the check, explain each result, a fix table for Audacity, Audition and Reaper users, and how the measurements are made
    • scripts/acx_check.py (main CLI), audio_io.py (pure-Python WAV and MP3 frame readers), make_samples.py (passing and failing test files)
    • references/acx-requirements.md: every requirement with its source on acx.com
    • references/measurement.md: exactly how RMS, peak, noise floor and room tone are measured, how MP3 headers are read, and validation results

    Standard-library Python only. No audio libraries, no API keys, no network calls, no installs.

    • MP3 audio isn't decoded (Python's standard library has no MP3 decoder), so RMS, peak and noise are measured on WAV files. Check the WAV master, or decode the MP3 to WAV with your own tool and check that.
    • Peak is sample peak, not oversampled true peak.
    • ACX doesn't publish the exact windowing its QA uses. Results within about 0.5 dB of a limit should be confirmed in ACX Audio Lab.
    • Checks technical specs only, not performance, pronunciation, edits, breaths, mouth clicks or plosives.
    • Doesn't check credits, retail sample content or chapter structure. The SKILL.md reminds you of those.
    • WAV files over 1 GB (and RF64) aren't supported. ACX's per-file limits are well below that.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    File Scopes

    **/*.wav
    **/*.mp3
    ./*.md
    ./*.csv
    ./samples/*

    All measurement is local and read-only. The scripts don't import networking or subprocess modules and never run external programs. Your audio stays on your machine.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    No. ACX's own analyzer is the final word. This catches problems earlier, on every file at once, with the fix for each, so you upload once.

    MP3 encoding can raise peaks. In our test, a master peaking at -4.0 dB measured -3.2 dB after a 192 kbps LAME export was decoded. Aim for -3.5 dB or lower on the master.

    It's the quietest half-second stretch in the file, ignoring stretches of exact digital zero, which are reported separately because ACX wants real room tone. Details are in `references/measurement.md`.

    No. The WAV and MP3 readers are written in plain Python. You only need your own editor to make the fixes.

    Yes. Pass a folder. It scans subfolders for `.wav` and `.mp3` and adds project-wide checks: mono/stereo consistency and loudness spread across chapters.

    New in Digital Products