Errors and limits
Error codes, HTTP statuses and API rate limits.
Errors come back as JSON with a stable error code, a readable message and sometimes details. Over MCP the same body is returned as a tool result with isError: true.
{ "error": "SAFETY_BLOCKED", "message": "This person was already contacted by another sender.", "details": { "reason": "contacted_by_other_sender" } }| Code | HTTP | Meaning |
|---|---|---|
BAD_REQUEST | 400 | The input failed validation. |
UNAUTHORIZED | 401 | Missing or invalid API key. |
PAYMENT_REQUIRED | 402 | Your plan doesn't cover this (for example, more senders than your trial allows). |
FORBIDDEN | 403 | The key can't access this workspace or resource. |
NOT_FOUND | 404 | No such action, sender, thread or contact. |
CONFLICT | 409 | The request clashes with the current state. |
SAFETY_BLOCKED | 422 | The safety layer refused the write. details.reason says why. |
RATE_LIMITED | 429 | Too many requests. Wait Retry-After seconds. |
UNAVAILABLE | 503 | A dependency is down. Try again later. |
Rate limits
Each API key can make 120 requests per minute across MCP and REST. This limits API calls only; how fast LinkedIn actions run is set by the safety rules.