Execute & Traces
Execute and traces
Drive the system under test and capture traces for assurance.
Execution produces the conversations or agent trajectories that assurance scores. Tracing captures the evidence; fetch normalizes it into one schema.
In this section
| Page | Use when |
|---|---|
| Recipe chatbot SUT | Local reference bot for tutorials |
| Tracing | Langfuse-centric capture today |
| Fetch tracing | Normalize traces into the assure schema |
Modes
- Active drive — platform/SDK sends messages or tool calls to the SUT
- Trace-only — SUT already runs in production; skip drive and proceed to fetch