get402 / policy

The get402 fetcher

For site owners who found get402-fetch in their logs.

What it is

get402 is a paid API. A caller, usually an AI agent, names one public URL and pays for the call. We fetch that one page, take out its main text, and return it as markdown. We do not crawl, follow links, or fetch pages nobody asked for.

How to recognize it

Every request we send carries this user agent:

get402-fetch/0.1.0 (+https://get402.ai/policy)

Requests come from cloud hosting addresses that change. Match on the user agent, not on IP addresses.

How to block it

We read your robots.txt and reuse the answer for up to 1 hour. We obey the group for get402-fetch. If there is no such group, the * group applies. To block us from the whole site:

User-agent: get402-fetch
Disallow: /

A blocked request is never fetched. The caller gets an error and is not charged.

What a request looks like

• One GET for the page, plus one for /robots.txt when we have not read it in the last 1 hour.
• At most 60 page requests to one host per minute, counting every caller together. Requests past that are refused before they reach you.
• At most 5 redirects. A redirect to a private or internal address is refused.
• We stop reading after 1,000,000 bytes and give up after 20 seconds.
• We keep the extracted text for up to 1 hour so a repeat request for the same URL does not hit your server again. We keep nothing else from the page.

Contact

Problems, questions, or a request we should not be making: contact@get402.ai.