Your agents don’t need to persist. Your purpose does.

How do I connect GA4 to ChatGPT?

The official GA4 server is read-only and client-agnostic, so ChatGPT gets the same traffic read.

Direct answer: connect the official GA4 MCP server (read-only) to ChatGPT. In ChatGPT, add it from the connector settings (developer mode); the tool's own docs carry the current path. Authorise your property and the read covers sessions, users, events, conversions, engagement, sources, landing pages, device and geo, and e-commerce revenue. Orientation reads are fine; confirm decision-grade numbers in the GA4 interface, as you would with any report.

Orientation, honestly priced

Which pages drew traffic, how sources shifted, what the conversion trend looks like — quick reads with real convenience, read with the care any report deserves. Unl judges the data it is given; checking the source stays your job either way.

Attribution is a treaty, not a truth

Worked case: your team set an attribution ruling — decisions cite the 30-day window, full stop, because a channel argument once ran three meetings on numbers from three different windows. GA4 computes any window; that one of them is decision-grade in your company is a treaty you signed with yourselves.

Through Unl the channel read arrives under the treaty: these are 30-day figures — the citable ones; other windows arrive labelled as context, not currency. The three-meeting argument retires.

GA4 to ChatGPT is a read-only official server; through Unl the traffic answers to the attribution treaty you set — one window is currency, the rest is context.

Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. It works today: 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 the GA4 server read into ChatGPT?

Sessions, users, events, conversions, engagement, sources, landing pages, device/geo and e-commerce revenue — read-only, from the authorised property.

How should I treat the figures day to day?

Use conversational reads for orientation and pattern-spotting; before money moves on a number, confirm it in the GA4 interface.

Can a ruling fix inaccurate data?

No — rulings govern judgement, not data quality. The attribution treaty ends window arguments; the figures themselves deserve the same verification whatever lane they arrive through.

What if two windows disagree wildly on the same campaign?

Then the treaty earns its keep: one window is the one you spend against and the rest are context, decided in advance precisely so this moment does not become a negotiation. The disagreement is still worth reading. It simply does not get to reopen which window is currency.

What this is

Your agents don’t need to persist. Your purpose does.

Save the thoughts, decisions and targets worth keeping, each with its reasoning, carried into every AI session the moment they matter. Unprompted.

Your whole AI world. What you decided at the epicentre. It works inside Claude, Claude Code, ChatGPT and Cursor, and connects in a couple of steps.

MCP native·You decide·Model agnostic·Your data

The right information

Connect Unl to bring the right information into the moment.

Your sources, read against the criteria you set.

Add Unl to your AI

The full product, open. Free to start. Paid plans from $10 a month.