Remote Endpoints
Remote Endpoints let you connect external or self-hosted model servers to SecureAI through AI Gateway controls. This is useful when your organization runs its own model infrastructure and still wants centralized security governance.Where To Configure
Go to Admin → AI Gateway → Remote Endpoints.Registration Options
- Quick Install: guided wizard that generates a one-command setup flow.
- Manual Registration:
- Paste JSON
- Manual Entry fields
Connectivity Modes
- Reverse Connect: recommended for most deployments; outbound connection, no open inbound ports.
- Cloudflare Tunnel: zero-trust connectivity using tunnel token and hostname.
Daily Operations
For each endpoint, admins can:- run Health Check
- run Verify mTLS
- inspect endpoint details (region, fingerprint, models, last check)
- remove endpoint when decommissioned
Suggested Onboarding Flow
- Start with Quick Install.
- Choose connectivity mode.
- Select required models.
- Complete registration.
- Run Health Check and Verify mTLS.
- Confirm endpoint remains healthy after 5 to 10 minutes.
Troubleshooting Checklist
- Endpoint unreachable: verify connectivity mode and hostname values.
- TLS passed but unhealthy: check remote gateway/model service status.
- Not connected: confirm remote service is running and registration is valid.
Best Practices
- Prefer Reverse Connect unless your architecture requires tunnel mode.
- Keep endpoint ownership clear by organization/team.
- Re-check endpoint health after policy or key changes.