Press kit

The system of record for AI agent decisions.

Every AI agent in production makes decisions that companies will have to prove. Runback is the first platform that makes any decision re-executable, auditable, and signed.

press@runback.dev →Security & trust →

Founded: 2025Headquartered: Melbourne, Victoria, AustraliaLegal entity: Runback Pty LtdCategory: AI agent governance — observe, replay, gate, auditDeployment: Managed cloud · Self-hosted in your VPCFrameworks mapped: EU AI Act Art. 12 · APRA CPS 230 · NIST AI RMF · ISO/IEC 42001Certifications held: None — no SOC 2 or ISO 42001 report, none in auditPricing: Free self-hosted · Starter A$49/mo · Growth A$600/mo · Scale A$1,800/mo · Pro A$5,000/mo · Enterprise customPress contact: press@runback.dev

Boilerplate

One paragraph.

Runback is the system of record for AI agent decisions — the first platform that captures every model call, tool use, and reasoning step, then re-executes any of them deterministically from the exact captured context. It is built to reproduce incidents in under 5 minutes, gate every model or prompt change in CI, and export a tamper-evident audit artifact mapping to EU AI Act, APRA CPS 230, and NIST AI RMF requirements. The platform runs fully inside the customer's perimeter on Enterprise.

Verifiable claims

Check everything we say.

3h 48m → 4m 23s MTTR

A worked reference scenario, not a customer case study — we're too early to have one. Loan-approval agent, run_a3f1b90c, same mechanism a real incident uses. Walk it live at runback.dev/how-it-works.

6-layer determinism substrate, proven in public CI

The replay oracle's proof run is public — inspect it yourself. See the actions ↗

EU AI Act claim

Article 12 is in force. Read it ↗

Press contact.

Embargoed briefings, interview requests, or brand assets — press@runback.dev. One business day response.

press@runback.dev →Security & trust