More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Mastering Engineer: Streaming Loudness

    1

    Finish approved mixes for release: measure first, run the ordered mastering chain, hit streaming loudness and true-peak targets, and pass real QC checks.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Mastering Engineer: Streaming Loudness

    Mastering Engineer: Streaming Loudness

    Example session with this skill installed

    Give me a mastering plan for an indie-folk track. Measured mix: integrated −19.5 LUFS, true peak −0.2 dBTP, LRA 11.5, 24-bit/44.1 kHz WAV. Releasing on Spotify and Apple Music.

    • Read your context and instructions
    • Compiled the mastering engineer

    Diagnosis. −19.5 LUFS needs lift, but folk is a genre exception: dynamics are the point, and Apple Music normalizes to −16. Decision: master to

    −16 LUFS, true peak −1.0 dBTP. Apple plays it natively; Spotify turns it up ~2 dB, which is harmless — unlike turning a crushed master down. LRA 11.5 is wide but right for folk; the job is to protect it, not tame it. True peak −0.2 is hot: the limiter ceiling handles it, but verify no clipping exists in the mix itself first.

    Chain.

    1. Trim head/tail; 10 ms fade in, natural fade out — folk endings breathe, don't chop them.
    2. High-pass at 20 Hz (acoustic material setting, not 30) to clear rumble without touching body.
    3. Corrective EQ: listen for fingerstyle harshness around 3–5 kHz; cut only if present, ≤2 dB, narrow. Nothing pre-committed.
    4. Dynamics: bus compression 1.5:1, slow attack, under 1 dB of reduction — glue only. Multiband stays off; the LRA is a feature.
    5. Tonal EQ: at most a +0.5 dB high shelf for air if the QC listen says the top is dull.
    6. Stereo: mono fold-down check (much folk listening is one speaker); bass stays centered.
    7. Limit: ceiling −1.0 dBTP, gain up ~3.5 dB until peaks kiss 1–2 dB of reduction. Stop there — the remaining loudness comes from the platform's turn-up, not more limiting.
    8. No dither: delivery stays 24-bit.

    Expected result: −16.0 LUFS ±0.3, TP −1.0 dBTP, LRA ≈ 11. Then QC in order: codec preview, mono, phone speaker (vocal must survive), quiet-volume listen — folk falls apart at low volume if the midrange is thin, and that's the check that matters most here.

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

    About this skill

    Mastering is the last audio stage before a track leaves the studio, and most of its failures come from skipping the measurement and reaching for the limiter. This skill is a tool-agnostic mastering method — the principles hold in a DAW, a mastering suite, or with Python DSP libraries (a runnable measurement snippet using pyloudnorm and soundfile is included). It starts with numbers, not ears: integrated LUFS, true peak, loudness range, and hygiene checks, with the red flags called out. Then it runs the chain in its load-bearing order — top-and-tail trim, sub-bass cleanup, corrective EQ, gentle dynamics, tonal EQ, stereo verification, limiting to a −1.0 dBTP ceiling, and dither only where bit depth actually drops. Delivery targets are tabled per platform (Spotify and YouTube at −14 LUFS, Apple Music at −16, with genre exceptions for acoustic, jazz, and classical work), along with the reason pushing louder gains nothing under normalization. A "fix here or send back" decision table routes part-level problems (a buried vocal, one harsh stem) back to the mix stage instead of papering over them, and edge-case guidance covers mixes that arrive already limited, clipped sources, spoken word, mono material, and re-masters. Album cohesion rules and a five-part QC routine (codec preview, mono, phone speaker, three volumes, second system) close the job, and every delivery includes a before/after measurement sheet so the result is verifiable.

    • It cannot fix mix problems — balance, arrangement, and stem levels are routed back to the mix stage by design.
    • A mix that arrives already limited or clipped caps what is achievable; the skill documents the constraint rather than undoing it.
    • Platform normalization figures are current published targets; platforms can and do adjust them, so targets should be re-verified for major releases.
    • QC verdicts depend on the user's playback systems — the routine specifies at least two systems plus a codec preview, but it cannot listen for you.

    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 24 days ago

    Permissions

    No special permissions declared or detected

    Listed24 days ago
    Updated21 days ago

    What's inside

    Frequently Asked Questions

    Finish approved mixes for release: measure first, run the ordered mastering chain, hit streaming loudness and true-peak targets, and pass real QC checks.

    This skill uses the open SKILL.md standard. Use it through an MCP-compatible agent connected to Agensi, and check the compatibility information on this page for confirmed agents and required capabilities.

    Connect your AI agent to Agensi through MCP, then ask it to use Mastering Engineer: Streaming Loudness on your task. Your agent loads the latest approved version each time. Paid skills require a purchase or active subscription. Skill files are not available for download; your licence does not allow you to save, copy or export any part of a skill.