Planck.AI.Models.Ollama (Planck.AI v0.1.0)

Factory for Ollama models via its OpenAI-compatible HTTP server.

Like llama.cpp, the available models depend on what has been pulled into the local Ollama instance, so this module provides a factory rather than a static catalog.

Examples

iex> Planck.AI.Models.Ollama.model("llama3.2")
%Planck.AI.Model{provider: :ollama, base_url: "http://localhost:11434", ...}

iex> Planck.AI.Models.Ollama.model("qwen2.5-coder:7b",
...>   base_url: "http://10.0.0.5:11434",
...>   context_window: 32_768
...> )

Summary

Functions

model(id, opts \\ [])

Builds a Planck.AI.Model for an Ollama-hosted model.

Functions

model(id, opts \\ [])

@spec model(
  String.t(),
  keyword()
) :: Planck.AI.Model.t()

Builds a Planck.AI.Model for an Ollama-hosted model.

Options

:base_url — base URL of the Ollama server. Defaults to http://localhost:11434.
:context_window — context window size. Defaults to 4096.
:max_tokens — max tokens to generate. Defaults to 2048.
:supports_thinking — whether the model supports thinking blocks. Defaults to false.
:input_types — list of supported input modalities. Defaults to [:text].
:default_opts — inference parameters applied on every call unless overridden by the caller (e.g. [temperature: 0.8, top_p: 0.9]). Defaults to [].