AI Runtime Assurance
Start with one high-risk AI workflow.
Book a focused Agent Runtime Security & Evidence Sprint, then deploy runtime assurance where the risk is real.
Agent Runtime Security & Evidence Sprint
Start with one named workflow. Map the agent surface, identify runtime control gaps, prove the evidence path, then deploy runtime assurance where the risk is real.
Workflow and drift readiness
The Sprint maps model calls, tool calls, drift signals, escalation paths, and operating events so your team knows where runtime assurance should be instrumented first.
Runtime controls
Deployed in your environment. Local controls allow, block, redact, restrict, escalate, or require review before risky behavior reaches a workflow, tool, record, or customer.
Security-review readiness
Focus the evidence on the buyer question: what the agent can do, what controls exist, what ran, what was blocked or escalated, and what remains to improve.
Zero sensitive-data egress
Glacis runs inside your infrastructure. Controls execute locally. Prompts, outputs, PHI, customer data, source code, credentials, and proprietary context stay inside your stack. Glacis exports only verification metadata, signatures, hashes, and evidence artifacts designed to prove control execution without exposing sensitive payloads.
Evidence accumulates
Every control decision, drift signal, escalation, and signed receipt becomes review-ready evidence: evidence packs, security reviews, regulatory artifacts, and learning for the next control update.
Pricing built around the first risky workflow
Agent Runtime Security Assessment / Sprint
A 10-business-day sprint on one AI workflow: agent/tool surface, runtime controls, evidence gaps, security-review readiness, and customer-ready proof.
Runtime Assurance Platform
Local runtime controls, signed receipts, evidence packs, policy mapping, and zero sensitive-data egress.
Evidence Packs
Review-ready artifacts for enterprise security reviews, audits, regulated clinical AI evidence, insurance, and customer trust.
Founder-design-partner pricing is available for the first three customers — one named workflow, 10-business-day scope, 100% upfront, anonymized case-study and reference permission. Ask us on the call.
Not sure which tier fits? Grab 25 minutes with us — we’ll scope the workflow and evidence path.
Signed proof that runtime controls executed.
Runtime controls execute locally and generate signed receipts written to the OVERT 1.0 open standard. Receipts chain, tampering is detectable, and third parties can verify control execution without seeing the sensitive payload.
- • OVERT-conformant signed receipts, assembled into evidence packs
- • Framework exports: NIST AI RMF, ISO 42001, EU AI Act, OSCAL
- • Zero sensitive-data egress — evidence packs assembled from local receipts
Pricing FAQ
What does the Sprint cover?
One named AI workflow over 10 business days: agent and tool surface, runtime controls, evidence gaps, security-review readiness, and a customer-ready proof artifact.
What happens after the Sprint?
Teams usually deploy runtime assurance where the risk is real: local runtime controls, signed receipts, evidence packs, and policy mapping for the workflow that needs enterprise trust.
We already use Vanta/Drata. Do we need GLACIS?
Vanta and Drata document that you have policies. GLACIS proves you followed them at runtime. They’re complementary — together they close the AI evidence gap.
Bring us one AI workflow.
We’ll map the agent surface, identify the runtime control gaps, and show what proof your customers will expect before they trust it.