Skip to content

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.