Runback
Blog

Engineering notes and governance explainers.

What we're proving, how the replay engine actually works, and what it means for teams that need to show their work.

Subscribe via RSS →
integrationsThree ways to wire an agent into RunbackHow to pick between the Vercel AI SDK integration, an OpenTelemetry exporter, and the framework-agnostic manual recorder — with the actual commands for each.Runback Team · July 14, 2026engineeringWhy replay-from-step-N is harder than it looksWhy capturing the request at the wrapLanguageModel boundary — not at your own call sites — is what makes replay mean something.Runback Team · July 14, 2026securityRedacting secrets before they ever leave your processHow @runback/redact scrubs API keys, JWTs, SSNs, and card numbers in-process, before anything is sent anywhere.Runback Team · July 14, 2026complianceWhat a hash-chained audit trail actually provesThe exact algorithm behind Runback's signed, append-only audit record — and what it means for EU AI Act and APRA CPS 230 compliance.Runback Team · July 14, 2026productDebugging an AI agent with time-travel replayA walkthrough of a real failing agent — from error-first navigation to editing the exact captured request and replaying it.Runback Team · July 14, 2026engineeringWe deterministically reproduced a real data raceA stress test of Runback's replay engine: reproducing a genuine, lock-free data race byte-for-byte, three runs in a row, proven in CI on every push.Runback Team · July 14, 2026engineeringBisecting a model regression in 4 probes instead of 12git bisect for agent behaviour: binary-searching a candidate timeline for the exact model or prompt change that introduced a regression, in O(log n) re-executions.Runback Team · May 27, 2026engineeringHow we auto-mine production incidents into regression testsA failed run becomes a permanent golden-suite entry, deduped by signature so a thousand identical incidents collapse into one test that keeps checking itself.Runback Team · April 2, 2026productSimulate a policy against history before you enforce it liveHow Runback's policy simulator checks a candidate rule against 100 recorded runs — no model calls, no replay — and names the exact runs it would have blocked.Runback Team · February 18, 2026