Limits

Over a limit you get a 4xx error that names the limit and the value. We never quietly lower quality or resolution.

Per plan

Plan	Renders a month	Renders a minute	Cache hits a minute	In flight at once
Free	50	10	600	1
Hobby	1,000	20	600	1
Pro	10,000	60	600	2
Scale	50,000	120	600	2

Rates are per API key; in flight is per account. Months are UTC calendar months.

Render time	20 s per render
Render slots	2 renders at once across the whole service, 8 waiting, a wait of at most 10 s
Request body	1 MB
OG html	256 KB
OG css	64 KB
PDF html	512 KB
PDF header_html and footer_html	16 KB each
Requests a page may make	300 requests per render (the page renders what it has, and X-Render-Warnings says so)
Size of one fetched response	10 MB
Total fetched per render	50 MB
Redirects	5 redirects per navigation
Allowed ports	80, 443, 8080 and 8443
Allowed schemes	http and https (data: and blob: inside a page work without the network)
Full-page screenshot height	16,384 CSS px (longer pages are cut and flagged with X-Render-Truncated: true)
Screenshot width	viewport width times device_scale_factor, at most 7,680; a wider full-page capture is cut at the right edge and flagged with X-Render-Truncated: true
Image pixels	32 Mpx: width times height after device_scale_factor, at most 32,000,000 pixels (1920 x 16,384 or 3840 x 8,192 at factor 1 fit). A bigger capture is refused with output_too_large, naming the limit and your size
Image output	10 MB
PDF output	20 MB and 200 pages
hide_selectors	20 selectors, 256 characters each
cache_ttl	0 to 86,400 s (1 day); the cache holds at most 2 GiB in total
delay_ms	0 to 10,000
Login links	3 per email per hour, 10 per network per day; a link works once, for 15 minutes