You got here from an MCP tool call
Your agent just converted a query. KeyKit is what tells you if the provider on the other end actually holds up.
Right now you are wiring up a provider, and the query converter solved the syntax problem. It cannot tell you whether that provider actually meets your requirements, coverage, freshness, reliability, and the rest. That is what an evaluation does, run through your own trial key, against tests tied to what you actually need.
Start an assessment