---
title: Rate limits
description: API rate limits and how to work with them.
---
import { Callout } from "fumadocs-ui/components/callout";
Docfork enforces rate limits to keep the API stable for all users. If you send many requests in quick succession, you may receive `429` responses. Authenticated requests use higher limits than anonymous traffic.
## Limits by user type
| User type | Search | Regular | Monthly | Rate key |
| ---------------------------- | ------ | ------- | -------- | ------------ |
| **Anonymous** | 20/min | 120/min | 500/mo | IP address |
| **Authenticated** (API keys) | 60/min | 300/min | 1,000/mo | organization |
| **Marketing site** | 60/min | 300/min | None | End-user IP |
- **Search** — `/v1/search` and MCP `query_docs` calls. More expensive to run.
- **Regular** — Other API routes (like `/v1/read`) and MCP `fetch_url`.
- **Monthly** — Rolling 30-day window. Resets continuously.
Rate limits are enforced per organization for authenticated requests, so all API keys in the same workspace share the same limits.
<Callout type="info" title="Higher limits">
Need more? [Contact us](mailto:support@docfork.com) — plan upgrades and higher limits are coming
soon.
</Callout>
## Response headers
Successful responses include rate limit headers:
| Header | Description |
| ------------------------------- | ----------------------------------------- |
| `RateLimit-Limit` | Max requests per window |
| `RateLimit-Remaining` | Requests left in current window |
| `RateLimit-Reset` | Unix timestamp when the window resets |
| `X-Monthly-RateLimit-Limit` | Monthly limit (when applicable) |
| `X-Monthly-RateLimit-Remaining` | Monthly requests left |
| `X-Monthly-RateLimit-Reset` | Unix timestamp when monthly window resets |
When a limit is exceeded, responses include `Retry-After` (seconds until reset).
## Rate-limited responses
Requests that exceed a limit return `429 Too Many Requests` with a JSON body:
```json
{
"error": "Too Many Requests",
"message": "Rate limit exceeded. Please try again later."
}
```
For anonymous search limits:
```json
{
"error": "Too Many Requests",
"message": "Search is limited to 20/min for anonymous users. Create a free account for higher limits at docfork.com."
}
```
For monthly limits:
```json
{
"error": "Too Many Requests",
"message": "Monthly rate limit exceeded. Create a free account at docfork.com for higher limits."
}
```
## Handling rate limits
1. **Watch for `429`** — Implement retries when you receive this status.
2. **Use `Retry-After`** — Wait at least this many seconds before retrying.
3. **Exponential backoff** — Add randomness to avoid thundering herd when many clients retry at once.
4. **Authenticate** — Use an API key for higher limits. See [Authentication](/core/authentication).
For MCP clients, consider a token-bucket style limiter on the client to stay under limits.
## Anonymous vs authenticated
Anonymous requests (no `Authorization` or `DOCFORK_API_KEY` header) use the lowest limits.
<Callout type="info" title="Get higher limits">
Add an API key for 3× higher per-minute limits and 2× higher monthly limits. [Create a free
account](https://app.docfork.com/signup) and add your key to the `DOCFORK_API_KEY` header.
</Callout>
| Limit | Anonymous | Authenticated |
| ------- | --------- | ------------- |
| Search | 20/min | 60/min |
| Regular | 120/min | 300/min |
| Monthly | 500/mo | 1,000/mo |
See [Usage](/core/usage) to track your workspace consumption.
