Skip to main content

Confirm

Are you sure?

Changelog · September 20, 2026

One credit meter, priced by what each call costs

Light work stops subsidising heavy work, and the bill says where the credits went.

By · Last updated: September 2026

TL;DR

Plans now include one monthly balance of credits instead of two separate allowances of AI and web calls. /html, /markdown, /screenshot and /console cost 1 credit, /lighthouse 2, /meta 5, /keywords 15 and /summarize 17. Cache hits and failed requests stay free.

Free plan, no credit card. 1,000 credits a month.

The old plans counted AI and web operations separately, one call each, as though every call in a bucket cost the same. They don't: a summary and a rendered HTML page differ by more than a thousandfold in what they cost to serve. One meter, with every endpoint priced by its cost, means you pay for the work you do.

CallCredits
/html, /markdown, /screenshot, /console1
/lighthouse2
/meta5
/keywords15
/summarize17
/scrapeThe sum of its operations

X-Quota-Cost on every response says what that request actually spent, and the 429 body gains needed. The organization screen shows which operations spent the most.

FAQ

Frequently asked questions

What does a cache hit cost?
Nothing. Cache hits and failed requests spend no credits.
Why does /markdown cost the same as /html?
Because it calls no model: it is a conversion of the rendered page, which costs one page visit.
How do I see what a request cost?
Read the X-Quota-Cost header on the response.

Try it on the free plan.

Free plan, no card. Confirm your email and your API key is live — you'll be making real requests in minutes.