The MCP connection is included in every SISTRIX package. MCP requests do not consume API credits, so there are no additional retrieval costs on the SISTRIX side.
Requests are subject to a fair use limit. It is set generously: normal work in the chat does not come close to it, and larger packages have a correspondingly higher allowance. In practice it is only reached with very heavy usage, for example when an assistant loops through entire keyword lists or complete projects, because every data point becomes its own request. The server then returns “Daily MCP quota exhausted” (error 7010), and the quota is available again the next day.
There is no live counter yet. However, the request log under Account > API lists every single MCP request with a timestamp, so you can follow your own usage. For regular larger data volumes, the classic API is the intended route, and its scope depends on your package.
On top of that, the AI platform applies its own limits: Claude and ChatGPT run on their own subscriptions or API usage and restrict usage independently of SISTRIX.
