The Glass Box
Every answer ships with its reasoning trace — the schema inspected, the queries run, and the results — so you can catch a wrong step early.
The Glass Box Principle
Transparency is core to canTrust. Every assistant response includes anAnalytical Reasoningtrace that shows the AI's internal logic.
What You Can See
- Thought Process — the agent's step-by-step reasoning phases.
- Tool Calls — the exact SQL or Python code generated and executed.
- Data Snippets — raw results retrieved from your sources before synthesis.
Click the reasoning indicator above any response to expand the trace and verify the logic.