Runtime assurance for AI systems that act.
The runtime proof layer underneath your AI — local classification, action-boundary enforcement, verifiable receipts.
- Probe adversarial testing
- Block runtime controls
- Sign signed receipts
Probe. Block. Sign.
Most teams can describe policies, logs, and evals; few can probe the boundary where the agent acts, block what shouldn’t happen, and produce a portable receipt of what did.
Probe — find what’s exposed
Authorized adversarial probing against APIs, tools, agents, and workflows inside your environment. Every finding lands with the request that proves it — exploitable, not theoretical.
Block — harden at the action boundary
Runtime controls sit inside your environment, with classifiers running on your own hardware. Tool allowlists, parameter checks, approval gates, regression tests in CI — PHI, secrets, and customer data never leave your stack to decide whether they should.
Sign — emit a verifiable receipt
Every probe, block, and controlled action emits an OVERT receipt: what control ran, what decision it made, which workflow it protected. Signed at runtime, tamper−evident, verifiable offline — independent of Glacis. Receipts roll up into evidence packs for security review, compliance, buyers, and insurers.
Receipts feed back into the next probe pass. Incidents, drift, near misses, and control outcomes strengthen policies, monitoring, and the regression tests that ran in CI. The trio is the motion; the loop is how the system gets harder over time.
We lived the problem
We built the “ship fast vs. prove it works” paradox at SwiftKey, Vektor Medical, and Cognoa.
Joe Braidwood
Co-Founder & CEO
Founding team at SwiftKey (acquired by Microsoft for $250M), scaled the keyboard AI to 300M+ devices pre-acquisition; the technology now ships on 1B+ devices globally. Chief Strategy Officer at Vektor Medical. 15+ years building AI products at scale.
Rohit Tatachar
Co-Founder & CTO
Nearly 19 years at Microsoft leading Azure infrastructure at billion-dollar scale. Most recently on the AI Foundry team building the infrastructure layer for enterprise AI. Architect of the GLACIS runtime kernel.
Jennifer Shannon, MD
Co-Founder & Chief Medical Officer
Board-certified psychiatrist with 15+ years clinical experience. Medical Director at Cognoa — helped develop Canvas Dx, the first FDA-authorized AI diagnostic for autism. CHAI Coalition member.
What we believe
Private by design
We architect systems so we never see your data. Trust through verification, not promises.
Built for builders
Accountability that slows you down isn’t accountability — it’s a tax. We optimize for both speed and rigor.
Prove, don’t promise
Attestations over assertions. Verifiable evidence over policy documents.
Trust you can defend
The evidence we produce belongs to your team — to hand to auditors, regulators, customers, and reviewers, and have them verify it independently.
If you’re trying to figure out where you stand with AI, we’d like to hear about it.
Get runtime coverage
Join us
We’re building the runtime assurance layer for the AI economy.