> For the complete documentation index, see [llms.txt](https://docs.adaptria.locaria.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.adaptria.locaria.com/mcp-server/limits.md).

# Limits

## Request body size

**Maximum: 10 MB per request.**

This limit applies to all tool calls, but most notably to `upload_file`, which accepts file content as a base64-encoded string. Because base64 encoding inflates binary data by roughly 33 %, the practical ceiling for the raw file being uploaded is approximately **7 MB**.

If you need to upload a file larger than 7 MB, use `generate_presigned_url` instead. That tool returns a presigned URL you can PUT the file bytes to directly, so the gateway never sees the file contents and the 10 MB body limit does not apply.

## Per-call timeout

**600 seconds at the load balancer.**

Individual tool calls return well within this limit under normal operating conditions. The timeout exists as a safety net for unusually slow downstream operations.

## Rate limits

Rate limits are inherited from the API gateway and apply to every MCP tool call:

* **Per IP address** (before authentication): **100 requests per minute**.
* **Per API credential** (after authentication): **60 requests per minute**, applied to all requests regardless of method.

When a rate limit is exceeded, the API returns `429`. The MCP server surfaces this as an error to the calling agent; there is no automatic client-side retry. Allow the agent to decide how to back off and retry. Read the rate-limit headers on each response rather than hard-coding these values, as the configured limits may vary by environment: `RateLimit-Limit`, `RateLimit-Remaining`, `RateLimit-Reset`, `Retry-After`.