ollama

package

v1.2.2 Latest Latest Go to latest Published: Apr 13, 2026 License: MIT Imports: 9 Imported by: 0

Details

Rendered for

Package ollama manages the Ollama process lifecycle for Cortex.

It provides health checking, automatic startup, model detection, and model pulling — all used by the TUI, CLI, and app initialization.

This section is empty.

This section is empty.

This section is empty.

type Manager struct {
	// contains filtered or unexported fields
}

Manager handles checking, starting, and provisioning Ollama.

func NewManager(baseURL string) *Manager

NewManager creates a new Ollama manager. If baseURL is empty, defaults to http://localhost:11434.

func (m *Manager) BaseURL() string

BaseURL returns the configured Ollama base URL.

func (m *Manager) EnsureRunning(ctx context.Context) error

EnsureRunning checks if Ollama is running, starts it if not, and waits for readiness.

func (m *Manager) HasModel(ctx context.Context, model string) (bool, error)

HasModel checks if a specific model is available in Ollama.

func (m *Manager) IsRunning(ctx context.Context) bool

IsRunning checks whether Ollama is responding at the configured URL.

func (m *Manager) PullModel(ctx context.Context, model string, progressFn func(string)) error

PullModel runs `ollama pull <model>` and reports progress via the callback. The progressFn is called with status strings; it may be nil.

func (m *Manager) Start(ctx context.Context) error

Start launches `ollama serve` as a background process. It does not wait for readiness — use WaitReady after Start.

func (m *Manager) StartedByUs() bool

StartedByUs returns true if this manager started the Ollama process.

func (m *Manager) WaitReady(ctx context.Context, timeout time.Duration) error

WaitReady polls the health endpoint until Ollama responds or timeout expires.