How abiosoft/colima works
Colima provisions and manages per-profile Linux VMs on macOS/Linux, then configures container runtimes and optional Kubernetes or Incus. The main workflow is CLI start → configuration → app orchestration → Lima VM and runtime setup; daemon-managed networking and mount notifications support the running instance. README also documents model-runner commands and lifecycle, SSH, status, and maintenance operations. Edges cite only visible calls or README-documented capabilities; unsampled implementation links are omitted.
Main entry points: app.go App lifecycle · start.go CLI commands · runner.go Model runners · lima.go Lima VM adapter
Components 27
CLI and profiles
- CLI commands
cmd/start.go - Command root
cmd/root/root.go - Profiles and config
config/configmanager/configmanager.go - Profile state
store/store.go
Instance orchestration
- App lifecycle
app/app.go - AI model commands
cmd/model.go - Model runners
model/runner.go - Docker model backend
model/docker.go - Ramalama backend
model/ramalama.go
VM and host integration
- Lima VM adapter
environment/vm/lima/lima.go - VM configuration
environment/vm/lima/yaml.go - Host integration
environment/host/host.go - Image downloads
util/downloader/http.go
Container runtimes
- Docker runtime
environment/container/docker/docker.go - Containerd runtime
environment/container/containerd/containerd.go - Incus runtime
environment/container/incus/incus.go - Kubernetes
environment/container/kubernetes/kubernetes.go
Background services
- Daemon manager
daemon/daemon.go - Host processes
daemon/process/process.go - Mount notifications
daemon/process/inotify/volumes.go - VM networking
daemon/process/vmnet/vmnet.go
Other components
- Colima user
- Docker client
- nerdctl client
- kubectl client
- Incus client
- Model registries
Connections 30
- Colima user → CLI commands: invokes
- CLI commands → Command root: registers with
- CLI commands → Profiles and config: loads and validates
cmd/start.go - CLI commands → App lifecycle: starts instance
cmd/start.go - App lifecycle → Lima VM adapter: starts VM
app/app.go - App lifecycle → Docker runtime: starts runtime
app/app.go - App lifecycle → Containerd runtime: selects runtime
app/app.go - App lifecycle → Incus runtime: selects runtime
app/app.go - App lifecycle → Kubernetes: detects Kubernetes
app/app.go - App lifecycle → Profile state: reads and writes
app/app.go - Lima VM adapter → VM configuration: builds VM config
environment/vm/lima/yaml.go - VM configuration → Daemon manager: uses network state
environment/vm/lima/yaml.go - CLI commands → AI model commands: offers model commands
- AI model commands → Model runners: selects runner
cmd/model.go - Model runners → Docker model backend: dispatches Docker
model/runner.go - Model runners → Ramalama backend: dispatches Ramalama
model/runner.go - Model runners → Lima VM adapter: runs guest commands
model/runner.go - Model registries → Model runners: supplies models
- Docker client → Docker runtime: uses runtime
- nerdctl client → Containerd runtime: uses runtime
- kubectl client → Kubernetes: uses cluster
- Incus client → Incus runtime: uses runtime
- CLI commands → Daemon manager: manages services
cmd/start.go - Daemon manager → Host processes: selects processes
daemon/daemon.go - Daemon manager → Mount notifications: registers watcher
daemon/daemon.go - Daemon manager → VM networking: registers networking
daemon/daemon.go - Mount notifications → Docker runtime: tracks volumes
daemon/process/inotify/volumes.go - Mount notifications → Containerd runtime: tracks volumes
daemon/process/inotify/volumes.go - Lima VM adapter → Host integration: uses host actions
- Lima VM adapter → Image downloads: obtains images