> ## Documentation Index
> Fetch the complete documentation index at: https://help.skillsdb.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Rate Limits

> Per-key hourly allowances in three buckets

<Info>
  **Quick Summary:** Requests are counted per key, per hour, in three buckets: reads, writes and bulk. Every response carries `RateLimit-*` headers so you always know where you stand.
</Info>

<Warning>
  **Current status: limits are measured but not yet enforced.** The numbers below are provisional; they will be finalized from real usage data and announced to all credential holders before enforcement is switched on. Build your integration against the headers now and enforcement will never surprise you.
</Warning>

## Buckets and provisional allowances

| Bucket | Counts                   | Provisional allowance |
| ------ | ------------------------ | --------------------- |
| reads  | Every GET                | 1,000 / hour          |
| writes | POST, PUT, PATCH, DELETE | 300 / hour            |
| bulk   | Bulk endpoints           | 20 / hour             |

A per-key allowance override is possible for legitimate high-volume integrations — talk to your SkillsDB contact.

## The headers

| Header                | Meaning                             |
| --------------------- | ----------------------------------- |
| `RateLimit-Limit`     | This bucket allowance               |
| `RateLimit-Remaining` | Requests left in the current window |
| `RateLimit-Reset`     | Seconds until the window resets     |

Once enforcement is on, exceeding an allowance returns 429 [rate-limit-exceeded](/api-docs/errors/rate-limit-exceeded) with `Retry-After`. Windows are fixed UTC hours.