Self-hosting / Uncloud
Moving a workload is only useful if people can still use it.
A creative studio keeps large files on a local NAS. An AI team owns its GPU server. A business moves a predictable workload from a cloud VM to colocated hardware. Each still needs dependable ways to reach applications, administer systems and connect to outside services.
Fibmesh addresses that connectivity layer. Choose where the workload runs, then give each connection the path it needs.
See billing and ERP deployments on existing servers →One workload, different connections
Keep administration private. Publish what needs an audience.
Private work and administration
Use Networks for approved staff, file access and server administration.
Explore Networks →A public application
Use Publish for supported HTTP apps, or Public IPs when clients need an IP address and service port.
Explore Public IPs →Several workloads
Give VMs or hosted services separate addresses with a routed block. Keep each service’s firewall and ownership explicit.
Explore Routed Subnets →A practical migration
Move one service before moving everything.
- 01 / Map dependencies
Record callers, databases, scheduled jobs, external allowlists and service names. Decide which connections should remain private.
- 02 / Connect the new host
Confirm the endpoint or gateway, address family, public destination and any fixed outgoing source.
- 03 / Validate the application
Test real client workflows, upload capacity and recovery. Migrate data and credentials using the application’s own procedures.
- 04 / Cut over deliberately
Update DNS and partner references with a rollback window. The old cloud provider’s IP does not automatically become a Fibmesh address.
What you still operate
The connection is one part of running a service.
You remain responsible for power, hardware, backups, software updates and application recovery. Existing broadband capacity and upload limits still apply. Fibmesh does not migrate data or turn one server into a highly available cluster.
For hybrid deployments, connect supported hosts across providers through Networks. Floating public-subnet orchestration and automatic failover are separate planned capabilities.
Plan a hybrid environment →Start with a real workload
Where would you rather run it?
Tell us where the application is today, where it should run and who needs to reach it.
Discuss your deployment →