> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.paradex.trade/api/general-information/rate-limits/common-questions/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.paradex.trade/_mcp/server.

# Common Questions

**What algorithm do you use for rate limits?**

* Leaky Bucket Algorithm rate limiting
* This technique uses a "token bucket" system based on a maximum number of available request tokens for a given timeframe. Each request consumes one token, and tokens are continuously refilled at a fixed rate. When all tokens are consumed, further requests are blocked until new tokens become available.
* Distributed across instances: the system coordinates token usage across all our servers to maintain the global limit.

**What if I hit a rate limit, how long is the cool down period?**

Since token refilling is a continuous process, there's no fixed cooldown period. As soon as a single token is refilled, you can make another request. The wait time depends on how many tokens you've used and the refill rate - it could be just milliseconds if you're slightly over the limit, or longer if you've exhausted your entire token bucket.

> **Rate limit example**
>
> If you have a 600 requests/minute limit:
>
> * Tokens will refill at \~10 tokens per second (1 token every 100ms)
> * If you have 0 tokens available you will wait \~100ms for 1 new token to be refilled
> * For continuous requests: You can make 1 request every 100ms sustainably
> * If you want to make multiple requests quickly you will need to wait longer to accumulate more tokens (e.g., 500ms to get 5 tokens)

**How long will an API query run before timing out?**

* All public and private API endpoints have a default timeout set to 30 seconds
* In the case of a timeout, the API will return a 503 error