Back to blog

Mozonic in ChatGPT: What You Can Ask, What Needs Pro, and What It Can't Do

Mozonic is a plugin in the ChatGPT Plugins directory, listed as "Analyze, fix & master mixes". Connect it to any Mozonic account — the free plan included — and ChatGPT can see the mixes you have uploaded, read the full analysis of each one and turn it into advice. Ask it to fix, master or render anything and you need Mozonic Pro. Two things it cannot do on any plan: take an audio file from the chat, and listen to your song. Everything the plugin does starts from a mix you uploaded in the Mozonic app.

That last point trips people up, so it goes first. ChatGPT itself cannot analyze audio — it can transcribe speech on paid plans, but it has no idea what a WAV's loudness, frequency balance or stereo image is. The plugin does not change that. What it gives ChatGPT is the measurements Mozonic already made: integrated loudness, dynamic range, crest factor, phase correlation, stereo width, low-mid mud, high-frequency harshness, a band-by-band energy breakdown and a mastering-readiness score, plus Mozonic's own list of issues and suggested actions. ChatGPT reads those numbers and talks you through them. It is a reader of analyses, not a listener.

Connecting takes a minute

[@portabletext/react] Unknown block type "undefined", specify a component for it in the `components.types` prop

The same Mozonic MCP server behind the plugin connects to Claude, Cursor and other assistants; the connection guide for those lives on the MCP docs page. This article sticks to ChatGPT.

On every plan: read, list, check

These are the requests a free account can make, and in practice they are the ones most people make first. In the plugin's first nine days as a free-to-connect tool, listing mixes and reading a mix's analysis were the two most-used actions by a wide margin.

[@portabletext/react] Unknown block type "undefined", specify a component for it in the `components.types` prop

The monthly analysis allowance is set by your plan — five on Free, thirty on Starter, one hundred and twenty on Pro — and it is spent when you upload in the app, not when ChatGPT reads the result. Reading is free and unlimited.

On Pro: fix, master, render, download, chat

The second listing prompt — "Run Mozonic's autofix on my mix Midnight Drive." — is where the plan line sits. Any of the requests below on a free or Starter account comes back as a polite refusal with an upgrade link; Mozonic Pro is $29 a month or $278 a year, and the moment the subscription is active the same request goes through.

[@portabletext/react] Unknown block type "undefined", specify a component for it in the `components.types` prop

What it cannot do, on any plan

[@portabletext/react] Unknown block type "undefined", specify a component for it in the `components.types` prop

A sensible first session

Upload one mix in the Mozonic app and wait for the analysis. Install the plugin, connect, and ask "Am I connected to Mozonic?" to confirm the account. Then "analyze my latest mix and tell me what needs fixing", followed by the follow-ups you would ask an engineer: what to cut, where, how much, what to check afterwards. Go back to your DAW, make the changes, bounce, upload the new version, and ask ChatGPT to compare the two. That loop — measure, change, measure again — is the whole value, and it runs on the free plan. Pro is for the day you want Mozonic to make the changes itself.

If you want the long version of why a language model needs measurements in the first place, read Can Claude analyze audio? — the answer is no, for the same reasons that apply to ChatGPT.

For the connection steps in Claude and Cursor, and the full tool reference, see the Mozonic MCP guide and the MCP docs page. For what the numbers themselves mean, start with why mixes sound muddy and what an AI mixing coach actually does.

Ready to try it: install Mozonic in ChatGPT, or upload a mix for a free analysis first and connect afterwards — the plugin will find it.

Start with the numbers: run a free mix analysis and ask ChatGPT what they mean.