Pricing
Start with a pilot. Price production around security and rollout.
Good pilot signals
The best pilots start with a real code, cloud, or database control problem.
An agent can change code, infrastructure, databases, or another privileged system.
Your team has clear policy or risk boundaries to encode before execution.
Logs alone are not enough for review, security, or compliance.
100/100
Blocked
0.065ms
Local p95
195ms
Live p95Worker edge auth/assertion, Cloudflare Access plus runtime-host round trip, then policy and halt response. Policy compute traced around 8-17ms; network is the floor.
88ms
DB branch p95
74ms
Halt p95
Beta pilot
Pilot access
Free
Available during beta for teams proving the runtime security loop on a real use case.
Python and TypeScript SDKs
One runtime-control loop across policy, enforcement, monitoring, and proof
Canonical release-agent proof flow across standard control, process-backed qualification, and Docker high-assurance modes
Governed database control with approved templates, scoped credentials, copy-on-write branches, masked result evidence, and review-required / blocked SQL lanes
Signed evidence packet export and operator review surfaces
Access-protected runtime host with PostgreSQL-backed policy, enforcement, monitoring, and proof services
Protected RunPod and Vast.ai GPU proof paths live today through the same Training lifecycle and evidence model
Enterprise
Production plan
Custom
Scoped around deployment shape, regulated requirements, access control, auditability, and support expectations.
Unlimited agents and production rollout planning
Organization, workspace, and role-model rollout planning
Append-only audit-log and evidence-export model
Input-sanitization, bounds-checking, and validation review
Tenant-isolation and data-boundary review
Compliance and audit packaging support
Sandbox and enclave deployment-shape review
Dedicated onboarding and architecture review
Priority support and response standards
Deployment planning for hosted or controlled environments