Smith::LLM::Ollama
Inherits Smith::LLM::Provider < Reference < Object
Constants
DEFAULT_HOST = "http://localhost:11434"
DEFAULT_MODEL = "gemma4:latest"
Constructors
new(host : String | Nil = nil, default_model : String = DEFAULT_MODEL, timeouts : Timeouts = Timeouts.default)
SourceInstance methods
complete(request : Request) : Response
Sourcecomplete_streaming(request : Request, &on_delta : String -> Nil) : Response
smith talks to Ollama through its OpenAI-compatible endpoint, so the stream is plain OpenAI-shaped SSE rather than Ollama's own NDJSON.
default_model
Sourcehost
Sourcename
Source