# Usage and limits

> How GetIntel meters MCP calls, the monthly read quota on each plan, what needs an active plan, and how to check usage with get_usage.

Source: https://getintel.ai/docs/mcp/usage-and-limits/

## Monthly read quota

Every tool call counts as one read against your account's monthly quota, except `get_usage`, which is free. The quota is shared with the [GetIntel API](/docs/api/errors-and-limits/): API requests and MCP tool calls come out of the same monthly number.

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

The quota resets on the 1st of each month. When it runs out, calls return: *"Monthly read limit reached (N, shared by MCP and the API). Resets YYYY-MM-DD. Upgrade for a higher limit."*

## Check your usage

Ask your agent: *"How many GetIntel MCP reads have I used this month?"* It calls `get_usage`:

```json
{
  "period": "month",
  "resets_on": "2026-10-01",
  "reads": { "used": 1240, "limit": 50000, "remaining": 48760 },
  "drafts": { "used": 0, "limit": 999, "remaining": 999 }
}
```

## Active plan required

The MCP server needs an active plan or a trial in progress. Without one, calls return: *"MCP access needs an active GetIntel plan."*

## Other limits

- **Time ranges:** tools with a period default to the last 7 days and accept up to 90.
- **`recheck_technical`:** at most once an hour.
- **Result sizes:** list tools accept a `limit`, for example `get_sources` up to 50 and `list_tasks` up to 100.

## Keeping usage low

- Start broad with `get_overview` instead of calling many tools separately.
- Ask about a named topic or rival rather than everything at once.
- Reuse what the agent already fetched in the same conversation.