Credits

A scrape costs credits. The price is a flat lookup on two fields of your request — not a formula, and not affected by page size or duration.

`render_js`	`proxy_type`	Credits
`false`	`datacenter`	1
`true`	`datacenter`	5
`false`	`premium`	10
`true`	`premium`	25

Because it depends only on fields you control, you can work out the cost of a run before you send it. The console shows the same number from the same table.

The cheapest meaningful saving is turning rendering off. If the HTML you want is in the first response, render_js: false is 5x cheaper on datacenter and 2.5x on premium — and faster.

What is charged

A run is charged when the target actually responded:

ok: true — you got the page.
blocked: true — you got an anti-bot challenge instead of the page. The request was made, the proxy was used, so it is charged.

What is free

Anything that failed on our side, before the target answered:

queue_timeout — no capacity came free in time
engine_unavailable, queue_unavailable — our infrastructure
insufficient_credits — refused before any work
transport failures where ok is false

This is why a retry loop is safe. Retrying a timeout does not bill you twice; only the attempt that reaches the site is charged.

Checking your balance

GET /analytics/summary with a login token reports usage and what is left. Running out returns insufficient_credits (HTTP 402) on the next scrape, which is not retryable — top up first.

Credits are granted per tenant. Ask us to top up while self-serve billing is still being built.