Direct answers about the product and its boundaries.

LumeGrid is secure execution infrastructure for AI work. It combines worker configuration, action-level policy, durable execution, compliance flows, tenant controls, and evidence that stays connected to the outcome.

No. Workers are one surface of the platform. The broader system includes model and tool routing, roles and permissions, autonomy policies, budgets and stop conditions, compliance and privacy operations, electronic signatures, enterprise-readiness controls, vertical workflows, developer surfaces, and evidence.

Supported roles can be configured with a name, role, mandate, cadence, capabilities, data boundary, authority policy, and owner. LumeGrid is not limited to a preset roster. The fully self-serve path for arbitrary tenant-created roles is still being verified.

New workers begin without action authority. An operator can keep proposal-only review or opt into bounded authority for defined action classes under an active operating charter. Requests outside that boundary return for review.

The operating model covers role and mandate, schedule, tool and data grants, tenant-level spend limits, action-level review, autonomy bands, escalation, expiry, pause, and stop conditions. Exact controls depend on the verified product surface.

Operational actions produce tamper-evident linked receipts. Signed approvals and document and evidence flows use cryptographic signatures. The product direction is to keep intent, identity, policy, approval, execution, and outcome connected.

The platform includes signed compliance flows, privacy-rights records for GDPR rectification and CCPA do-not-sell requests, enterprise-readiness controls such as MFA policy and IP allowlists, trust-center artifacts, security-review evidence, and vertical-specific compliance workflows. Availability varies by tenant surface and deployment.

No blanket certification claim is made on this site. LumeGrid can record control artifacts and security-review evidence, and some workflows are designed for regulated data boundaries, but tooling and framework alignment are not the same as an independent certification.

Managed, self-hosted, and isolated deployment patterns are part of the platform direction. The actual environment, data boundary, and outbound dependencies are agreed and verified with each design partner.

LumeGrid is in private development. Architecture and implemented capabilities exist across the repository, but the site does not describe a customer path as generally available until that path is verified end to end.

Public plans have not been published. Design-partner scope and commercial terms follow the workflow, control boundary, deployment environment, and acceptance criteria.

Use the contact form or email [email protected]. Do not send passwords, access tokens, onboarding links, signing links, or other secrets.