Understand the system before you operate it.
Start with the application and network model, then follow the real paths for deployment, runtime, service access, MCP, registry, CLI, and self-hosted nodes.
Get started
Getting Started
The platform model and the shortest path to a CPU workload on Tabbify Cloud.
Read →Quickstart
Build an OCI image from GitHub and deploy it asynchronously to one workload target.
Read →Core concepts
Tenants, applications, deployment targets, private networking, ingress, and data resources.
Read →Build apps
tabbify.toml
OCI build, CPU runtime, lifecycle, routes, environment, and one workload target.
Read →Runtime
How a CPU OCI image runs inside a Firecracker microVM.
Read →Workload lifecycle
Health gates, current lifecycle state, restarts, and bounded local log output.
Read →Services
Make HTTP calls between applications inside one tenant-private network.
Read →Networking
Platform
The Node API
Selected REST and MCP operations plus public workload ingress.
Read →Connect an MCP client
Authorize an MCP client and discover the operations currently exposed as tools.
Read →Registry
The authenticated OCI image store used by the deployment pipeline.
Read →Auth & join tokens
Credentials for control operations and tenant-scoped mesh admission.
Read →Deploy
Operate production
Managed operations
Current platform evidence and the human processes agreed separately for each engagement.
Read →Managed data
Early-access PostgreSQL, KV, and S3-compatible resources with explicit limits.
Read →Troubleshooting
A practical diagnostic order for deploy, runtime, network, data, and access failures.
Read →