Do prompts or customer data pass through Cygrid?

No. The MCP Gateway inspects tool calls, not content. Prompts and payloads never pass through Cygrid, and logs record each action and the person behind it, never the prompt.

LAUNCH WEEK · IT-SA NÜRNBERGMeet the founders at Hall 6, Stand 142 · 27–29 OctoberReserve your slot →

LAUNCH WEEK · IT-SA NÜRNBERGReserve your slot →

Solutions · Security & compliance

Governance says what should happen. Cygrid enforces what is allowed.

Cygrid audits every identity and its access, enforces your policies on every action, and generates the evidence for NIS2, DORA and the EU AI Act, continuously, from one live graph.

Book a demo

What auditors ask

Every question auditors ask, answered from the live graph

Every verdict is written to an immutable decision log that streams to your SIEM and produces the reports regulators ask for.

Auditor questionWhat Cygrid shows

01Who has access to what?

A complete view of every identity’s effective permissions, including user accounts, service accounts, OAuth clients and AI agents.

02Are our admin lists complete?

Cygrid finds shadow admins whose rights come through groups and never appear in the platform’s own admin list, and shows the path from user to group to role.

03Is anyone holding rights they never use?

Cygrid puts every granted permission next to what was actually used, and proposes the exact reduction, reversible in one step.

04Who approved it?

The delegation record for every action: the person, the scope they granted and the agent that executed it.

05Did the controls work?

Cygrid enforces what is allowed, and proves it: risky access revoked, downscoped or blocked, with timestamps and the policy that acted.

06Which actions were malicious?

Every event is audited and classified, so malicious or risky actions are flagged as they happen.

07Does it work with our security tools?

Cygrid connects to your SIEM and security tools, so you see and act on identity risk from one control plane.

08Are we ready for the next audit?

Evidence is generated continuously, so you’re audit-ready on any day, not only at audit time.

Use cases

From written policy to live rules

WHILE ACCESS EXISTS

Departed employee still an Atlassian admin

When someone is offboarded, Cygrid removes the Atlassian admin role they kept, with their content and seat intact.

BEFORE FIRST USE

Scoped admin instead of standing admin

Permanent GitHub or Workspace admin rights are removed. Admin access is granted on request, for a limited time.

AT RUNTIME

One security policy for every identity

Rules like ‘no external sharing of finance files’ apply the same way to people, service accounts and AI agents.

See real-world use cases

Security review

What your security team will ask

DATA PATH

Prompts never pass through Cygrid

The gateway inspects tool calls, not content. Logs record each action and the person behind it, never prompts. Policy decisions take milliseconds.

SECURITY POSTURE

Audited and tested

ISO 27001 audited, SOC 2 Type II in progress and an external penetration test passed. Connectors are read-only by default, with write access only for fixes you approve.

HOSTING AND CONTRACTS