# API errors and limits

> GetIntel API error codes, the monthly read budget shared with the MCP server, the per-minute rate limit, and how to handle each one.

Source: https://getintel.ai/docs/api/errors-and-limits/

## Errors

Errors return a status code and a JSON body with a machine-readable `code`:

```json
{ "error": { "code": "brand_required", "message": "This account has several brands: pass brand_id. See GET /brands." } }
```

| Status | Code | What to do |
| --- | --- | --- |
| 400 | `invalid_parameter` | A parameter has the wrong type, is out of range, or is missing. The message says which. |
| 400 | `brand_required` | The account has several brands. Pass `brand_id`. |
| 401 | `invalid_api_key` | The key is missing, mistyped or revoked. |
| 403 | `feature_not_in_plan` | The data is part of the Pro and Growth plans. |
| 403 | `plan_inactive` | The account's plan isn't active. |
| 404 | `brand_not_found` | The `brand_id` isn't on this account. |
| 404 | `market_not_found` | The `market` isn't a live market of this brand. See `GET /markets`. |
| 404 | `not_found` | The topic, task or other item doesn't exist. |
| 409 | `integration_not_connected` | The endpoint needs an integration, such as Search Console, that isn't connected. |
| 429 | `quota_exceeded` | The monthly read budget is used up. |
| 429 | `rate_limited` | More than 60 requests a minute. Wait for `Retry-After` seconds. |
| 502 | `upstream_error` | A service GetIntel depends on failed. Retry later. |

## Monthly read budget

Each data request counts as one read. The budget is **shared with the MCP server**: reads from your API scripts and from your coding agents come out of the same monthly number.

| Plan | Reads a month |
| --- | --- |
| Monitor | 10,000 |
| Pro | 50,000 |
| Growth | 200,000 |

`GET /usage` shows what's used and left, and when it resets (the first of the month). `/usage` and `/brands` don't count against the budget.

## Rate limit

Up to **60 requests a minute per key**. Over that, the API answers `429 rate_limited` with a `Retry-After` header.

## Freshness

The API reads the same data as the app, so it's as fresh as the app. AI answers update after the daily scan (9am IST), brand perception and the technical scan update weekly, and AI traffic every few minutes to every hour. See [Data freshness](/docs/academy/metrics/data-freshness/). There's no need to poll more often than once an hour.