AI agents, run by OpenAI.

What is ChatGPT agent? OpenAI's AI agent that browses for people. Signature checks coming soon.

Run by OpenAI.
Purpose: AI agents. A visit means a person handed a task to an AI agent, such as comparing products or filling in a form, and the agent is doing it on your site.
No single robots.txt name. Block it at your server instead.
Signs its requests. Signature checks coming soon in SenseGap.

At a glance, the facts in one table.

ChatGPT agent at a glance. Taken from SenseGap's open bot directory, the same data that sorts every visit in the dashboard.

ChatGPT agent facts
	ChatGPT agent
Operator	OpenAI
Purpose	AI agents
Recognized by	`Signature-Agent` header
robots.txt name	None
Verification	Signature (coming soon)
Documentation	None listed

Purpose, what it does and why it came.

What ChatGPT agent does. SenseGap sorts every visit by purpose. ChatGPT agent is in the AI agents group.

ChatGPT agent is OpenAI's AI agent that browses websites and takes actions on behalf of a person. It runs a real browser, so its user agent looks like an ordinary visitor's. OpenAI signs its requests with HTTP Message Signatures (Web Bot Auth) and sends https://chatgpt.com in the Signature-Agent header.

What a visit means

A visit means a person handed a task to an AI agent, such as comparing products or filling in a form, and the agent is doing it on your site.

Detection, the header SenseGap matches.

How SenseGap recognizes ChatGPT agent. By its Signature-Agent header. A separate check then decides whether the visit is real.

ChatGPT agent browses with a normal browser, so its User-Agent header looks like any visitor's. What gives it away is the Signature-Agent header that comes with its signed requests. SenseGap matches that header against the value below.

Header matched

Signature-Agent: "https://chatgpt.com"

Verification, how to tell a real visit from a fake.

ChatGPT agent can't be verified yet. It signs its requests. Checking those signatures is coming soon in SenseGap.

How we verify bots

What this means

ChatGPT agent uses a normal browser user agent but sends a Signature-Agent header of https://chatgpt.com with a cryptographic signature (Web Bot Auth). SenseGap recognizes it by that header. Checking the signature itself is coming soon, so until then these visits are marked Unverified.

robots.txt, rules you can copy.

How to block ChatGPT agent. It has no single robots.txt name, so a rule at your server or CDN is the dependable way.

robots.txt is a request, not a lock. To enforce a block, match the Signature-Agent header at your server or CDN. In nginx:

Enforce it at the server

# nginx, inside a server block
if ($http_signature_agent ~* "chatgpt\.com") {
    return 403;
}

SenseGap reads your robots.txt and shows, bot by bot, what you allow and which bots fetched pages you disallowed. See what you get.

Trade-off, what blocking changes.

Should you block ChatGPT agent? It depends on what you want from OpenAI. Here is what each choice means.

If you allow it
People who delegate tasks to the agent can complete them on your site.
If you block it
The agent cannot finish tasks on your site, so the person may finish them somewhere else.

OpenAI, every bot it runs.

More bots from OpenAI. OpenAI runs 4 bots in the directory, each with its own job.

FAQ, ChatGPT agent in short answers.

ChatGPT agent, answered plainly