Solutions
Enterprise
Your security team will ask four questions about any AI agent: what is it, what can it reach, what did it do, and can we prove it. We answer each with an artifact, not a questionnaire.
Capabilities
Available means proven on our lab clusters as of 2026-10-24, not deployed for customers. In development is being built; Planned is designed, not built. We have no customers yet and hold no certification or authorization.
Your cloud account, your cluster
Status: In development. One Talos cluster per customer, built from OpenTofu code in the account your contract names. The lab cluster is proven; customer cloud providers are being built.
Workload identity instead of API keys
Status: Available. Each agent is a SPIFFE identity; secrets are released to that identity only (lab).
Signed, cited task records
Status: Available. Every task names the sources it read and is signed; your auditors verify it with cosign, offline (lab, synthetic example).
Bring your own key
Status: In development. Key custody in your cloud KMS, your HSM, or split among your people (Shamir), chosen in the assurance profile.
Single sign-on and passkeys
Status: Planned. Keycloak, one realm per customer, with WebAuthn policy restricted to the authenticators you approve.
Agent catalog
Status: In development. Agents seeded from O*NET occupations; you tick the tasks you want. Browse the catalog.
We hold no certification or authorization; see the trust center for the current, generated control status.