Usage tokens

Instead of billing per request, we bill by how much data a request actually returns. Every response you get back costs some number of usage tokens, deducted from your key's monthly usage limit.

How a request is priced

• A response costs roughly 1 usage token per 500 bytes returned, rounded up, with a 1 token minimum. A single-row get is typically 1 token; a read or an endpoint like getGames that returns many rows costs more the more data it sends back.
• An error still costs a flat 1 token minimum once your request has passed authorization (for example, bad arguments to a valid table/function) -- so probing for valid arguments isn't free.
• A request that's denied for access reasons (missing/invalid API key, restricted table, or missing read/write/query access) costs nothing.

Response headers

Every response includes:

X-USAGE-REMAINING
X-USAGE-USED
X-REQUEST-USAGE

X-USAGE-REMAINING and X-USAGE-USED reflect your key's usage against its monthly limit after this request. X-REQUEST-USAGE is the cost of just this request. See Headers for the full list of headers.

Monthly usage limit

Each plan has a monthly usage limit that resets on the 1st of each month. Once a request would take you over your limit, that request is still let through -- so a single large request can put you over -- but every request after that is rejected with:

{"error": "usage limit reached"}

X-USAGE-REMAINING will read 0 on that response, and note it still arrives with a 200 status code — see Errors. View our plans to see the usage limit for each tier.

Free trial

The free trial is limited to 500 usage tokens per month, some numeric columns are randomized, and bulk read access is limited to organization and division -- every other table is get()-only on a trial key.