Skip to main content
Rate limits are set per organization, or per user for personal accounts. All API keys in the same organization or account share one limit, and the limit depends on your plan. Each request counts once, regardless of its parameters, such as the search depth or the number of URLs.

Get your API key

Create an API key to get started.

Requests per Minute by Plan

Need a higher limit? Contact us.

Endpoint Limits

The following endpoints have their own lower limit, which is the same on every plan. For example, an organization with a 1,500 RPM limit can send up to 1,500 search requests per minute, but no more than 100 crawl requests per minute.

Usage Endpoint Rate Limits

The usage endpoint has a separate rate limit of 10 requests per 10 minutes.

Handling Rate Limit Responses

When an API-key request exceeds a rate limit, the API returns 429 Too Many Requests. If the response includes Retry-After, wait the indicated number of seconds before retrying. For example:
The 60-second delay is an example; use the value returned in the response. Error messages can vary by the limit reached; use the HTTP status and header for retry handling rather than matching the message text. Keyless access uses a different error format and separate limits.

Flex Tier Capacity

On non-paying plans, search requests that set max_age_hours are served under a flex tier. While Tavily is under heavy load, these requests are refused with 434 Flex tier capacity exceeded. Retry shortly, or omit max_age_hours to be served from cache.
When using the REST API, include your API key in the request header so the request is counted against your organization’s limit. More information about plans and Pay as you go can be found here.