How an answer is produced

flowchart TD
  A["Question"] --> B["Agent configuration"]
  B --> C["Optional retrieval or tools"]
  C --> D["Model response"]
  D --> E["Optional citations, tables, charts, or files"]

Depending on the Agent and request, DataFact can show reasoning summaries, execution steps, tool cards, retrieved chunks, data tables, charts, markdown text, citations, clarification requests, or generated files. These blocks are conditional. Their absence is not automatically an error.

The interface renders returned execution information; it does not expose a guarantee that every internal process is complete, correct, or fully shown. Treat tool output and model narration as evidence to inspect, not proof by themselves.