# Capacity

> Default limits for the public deployment.

Canonical: https://keta1930.github.io/icml2026-mcp/en/docs/operations/capacity



| Limit                   |                        Default |
| ----------------------- | -----------------------------: |
| Active sessions         |                          4,096 |
| Search queue            |                            512 |
| DashScope calls         |                  64 concurrent |
| Milvus calls            |                  32 concurrent |
| Per IP                  | 600 requests/minute, burst 100 |
| Global application rate |          6,000 requests/minute |
| Provider rate           |         24,000 requests/minute |
| Provider input          |        1,000,000 tokens/minute |

These values are intended for roughly 25–40 simultaneous users with short search queries. Nginx, the application, and the provider quota are separate controls. Measure queue delay, provider latency, Milvus latency, SQLite lock time, and memory before increasing them.