Think inside your AI world.
How do I connect Figma to ChatGPT?
Design files as structure — screens, frames, components — readable by ChatGPT over the same MCP surface every client shares.
Direct answer: connect Figma's MCP surface to ChatGPT. In ChatGPT, add it from the connector settings (developer mode); the tool's own docs carry the current path. Authorised, ChatGPT reads design files and their structure — what screens exist, how flows connect, which components a frame actually uses.
Structure over screenshots, again
What changed in the onboarding file, whether the new screens use sanctioned components, where the flow forks — answered from structure. Exported PNGs stop being the medium of design conversation.
Accessibility is a floor, not a preference
Worked case: your team ratified an accessibility floor — AA contrast on every shipped screen, no debate per screen, after an accessibility review of the shipped product returned a list long enough to be embarrassing. Figma renders any contrast you pick; that AA is where debate ends is a standard you set once so it never gets relitigated per screen.
Through Unl the file read arrives with the floor held: two new screens carry text under AA contrast — below your floor, the embarrassing-list story attached. The standard applies at design time, where fixing is cheap.
Figma to ChatGPT reads design as structure; through Unl the structure answers to the floors you ratified — sub-AA contrast gets caught in the file, not in the audit.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. The reach lane is live: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
Read further
Questions people ask
What does ChatGPT read from Figma?
Design files and structure — screens, frames, flows, components — sufficient for review-against-standards and what-changed questions.
Why ratify design floors rather than rely on review?
Review is a person having a good day; a ratified floor is present at every read with its reasoning. They compose — the floor catches what the tired reviewer misses.
Does the rule work in Claude too?
Yes — rules in Unl are surface-independent. AA is AA whichever assistant reads the file.
What if the brand colour itself fails the floor?
Then the floor and the brand are in genuine conflict and somebody has to choose, which is a decision rather than a bug. Record which won and why. Otherwise the exception spreads by precedent, and within a year nobody can say whether the floor applies to anything.
What this is
Think inside your AI world — you stay in command
Save the thoughts, decisions and targets worth keeping, each with its reasoning, carried into every AI session the moment they matter. A new unit of exchange between you and your AI: the Settled Why with standing that travels. Unprompted.
Your whole AI world. What you decided at the epicentre. It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector — quick to connect, in a couple of steps.
MCP native·Human settled·Model agnostic·Your data
Measured Context
Connect Unl to bring the right information into the moment.
Your sources, read against the criteria you set.
The full product, open. Free at launch.