Hi everyone. I'm Aniq, a Network Engineering student at Universiti Malaysia Sabah. For my final year project I built ZFORGE, a management dashboard that puts OpenZiti in the hands of a small IT team.
Why: small businesses here mostly use MAC address binding or a flat VPN. MAC randomisation breaks the first, and the second gives every signed-in device the whole network. OpenZiti solves both, but the CLI and policy model are a lot for a one-person IT team.
What it does
- Device identities with a guided enrolment flow and emailed setup links
- Apps, access rules, posture checks and location rules, all point-and-click
- Live sessions with an end-session action
- Service cloaking checks, so office apps keep zero open inbound ports
- A hash-chained audit log, grouped security incidents and push alerts
Architecture: a Next.js dashboard and a Laravel API proxy in front of an OpenZiti 2.0.0 controller on AWS EC2. Office connectors are private routers in Docker, on the same Docker networks as the apps, so nothing on the office server is exposed. A port scan of the office VM shows nothing but the management link.
Something I'd love feedback on: with OIDC/JWT clients there are no API or edge sessions, so I build live sessions from /fabric/v1/circuits (using tags.clientId and tags.serviceId). Ending a session deletes the identity's circuits and calls POST /edge/management/v1/revocations with type IDENTITY. Is that the approach you'd recommend, or is there a better signal for "this device is connected"?
Launch video: https://youtu.be/gjXBzeBfLrI?si=HH_jWVRJxXpt2uK0
Live site: https://zforge.my
LinkedIn: www.linkedin.com/in/aniqsfd
Thanks to the OpenZiti team and community. Happy to answer any questions.

