Braintrust observes.
Runback re-executes.
Braintrust is an eval platform with a prompt playground and dataset management. It's built for evals before deploy, not governance after — no re-execution of production incidents, no policy enforcement, no audit trail.
| Capability | Braintrust | Runback | Note |
|---|---|---|---|
| Eval runner and dataset management | ✓ | ✓ | |
| Prompt playground and scoring | ✓ | partial | Runback scores via policy, not LLM judges |
| Human review UI | ✓ | ✗ | Braintrust speciality |
| Re-execute a production incident step-by-step | ✗ | ✓ | |
| Live policy enforcement — block before it runs | ✗ | ✓ | |
| Golden corpus auto-mined from production failures | ✗ | ✓ | |
| Signed, tamper-evident audit export | ✗ | ✓ | |
| Regulatory mappings (EU AI Act · CPS 230 · NIST) | ✗ | ✓ |
Most teams use both — Braintrustfor day-to-day trace browsing, Runback for incident replay, CI gating, and the compliance artifact. The two don't overlap until the moment an agent misbehaves in production and someone asks you to prove it.
Try the replay on a real incident.
No signup. Walk a failing production run step by step and see what the model saw.