VICIPanel applies tiered rate limits per MCP token to protect your tenant from runaway AI loops and to keep resource usage predictable.
- Standard tools (read-only, fast)
- 60 requests per minute per token
- Destructive tools (state-changing)
- 10 requests per minute per token
- Heavy compute (reports, exports)
- 5 requests per minute per token
- Response on limit
- HTTP 429 with Retry-After header in seconds
If you need higher limits, for example, running an agentic workflow that legitimately needs 500 queries per minute, contact us. We can raise limits on Enterprise tenants.