Limits

There are two limits. Both return 429 when reached, with a different code.

Limit	Amount	Applies to	Code
Rate limit	60 requests a minute	Each key	`RATE_LIMITED`
Monthly allowance	50,000 requests a calendar month	All live keys of one organisation, or of one researcher	`QUOTA_EXCEEDED`

The AI4M team can set a different rate limit for an independent researcher. Your own figure is shown in Settings → API access.

The rate limit

When a key makes more requests in a minute than it is allowed, the response includes a Retry-After header with the number of seconds to wait.

HTTP/1.1 429 Too Many Requests
Retry-After: 23

{
  "statusCode": 429,
  "code": "RATE_LIMITED",
  "message": "This key is limited to 60 requests a minute"
}

Wait that long, then try again.

The monthly allowance

Only requests made with live keys count. Test keys do not. The count resets on the first day of each month, and your usage so far is shown in Settings → API access.

{
  "statusCode": 429,
  "code": "QUOTA_EXCEEDED",
  "message": "Monthly request quota reached. Contact the AI4M team if you need more."
}

There is no Retry-After for this one: retrying will not help until the month ends. If you need a larger allowance, write to [email protected].

Staying well inside the limits

Call Periods to see updated_at, and only fetch scores again when it changes.
Store what you fetch. One request returns every LGA in a state; 37 requests return the whole country.
Fetch boundaries once. They do not change from month to month.