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

# QPS Limit

> Read the billing account's configured inference requests-per-second limit

Read the configured inference QPS (requests per second) limit for your key's billing account. Keys on the same account share this limit for each metered endpoint.

This read works with zero or negative balances. It does not spend credits or consume inference QPS. Responses use `Cache-Control: no-store`.

## Examples

```bash cURL theme={null}
curl https://text.external-api.pangram.com/limits \
  --header "x-api-key: $PANGRAM_API_KEY"
```

```python Python SDK (1.2.0+) theme={null}
from pangram import Pangram

client = Pangram()  # Reads PANGRAM_API_KEY.
limits = client.get_limits()
```

## Response

```json theme={null}
{"qps_limit": 2.5}
```

**qps\_limit** `number | null`

Configured inference QPS limit. Fractional values are preserved. `null` means no account QPS limit is configured; other service capacity and protection limits still apply.

This value describes the account's configured inference rate.

## Read limits and errors

This route shares the independent [account-read budget](https://docs.pangram.com/api-reference/balance.md#read-limits-and-errors) with balance and usage calls. The read budget is separate from the inference QPS value returned here. Respect `Retry-After` on `429`.

| Status | Meaning |
| - | - |
| `401` | Missing, malformed, revoked, inactive, or otherwise invalid key/account. |
| `429` | Account-read rate or concurrency limit; observe `Retry-After`. |
| `503` | Account lookup unavailable or timed out; retry the read later. |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.