At a glance
Plan volumes and prices are on the pricing page.
Brand fetches
One request to the Brand API, Brand Context API, or Transaction API is one brand fetch, drawn from a single shared balance. A brand fetch is counted when we looked the brand up. On the Brand API that means a200, a 304, or a 404, and on the Transaction API a 200 or a 404. On the Brand Context API, only a returned context is counted, and no error response is charged.
Errors on our side and rate-limit refusals are never counted, and neither are these:
- Requests for
brandfetch.comon the Brand API and Brand Context API, so you can test without spending quota. HEADprefetch requests.204answers tocachedOnlyrequests.- Images and screenshots linked from a Brand API response.
200.
Every response to a request made with an API key carries your plan quota in x-api-key-quota and your usage this month in x-api-key-approximate-usage. We email you at 80% of your quota.
Overage
On a paid plan, brand fetches beyond your quota are billed at the overage rate, up to the spending limit you set in the dashboard. Once that limit is reached, or when it is set to $0, requests draw on any prepaid credits you hold. When those run out, requests are refused with429 and x-bf-error: quota_exceeded until the quota resets.
Overage is not available on the free plan, and it does not apply to the Logo API or Brand Search API.
Free allowances
The Logo API and Brand Search API allowances are soft limits, so going over does not block you straight away. On the Logo API, we email you as you get close. Traffic far beyond the allowance can eventually be refused with429 and x-bf-error: quota_exceeded.
Brand Search usage counts every GET answered 2xx or 3xx, including answers from the edge cache. Browser preflight requests, refused requests, and requests without a valid client ID are not counted.
Throughput limits
Throughput limits protect the service from bursts. They are separate from your monthly quota.
A throughput block returns
429. Most carry no x-bf-error header, which is how you tell them apart from a quota refusal. The Brand Search per-client limit is the exception and sends x-bf-error: rate_limited. For search-as-you-type, debounce keystrokes to stay under both limits.
If you need more throughput, talk to sales.