@coopah/bentley-provider-ollama

Ollama LLM provider for Bentley. Run local models (Llama, Mistral, Gemma, etc.) with your Bentley agents.

Install

pnpm add @coopah/bentley-provider-ollama

Make sure Ollama is running locally or at the specified base URL.

Dependencies

• @coopah/bentley-core
• @ai-sdk/openai ^3.0.41 (Ollama exposes an OpenAI-compatible API)
• ai ^6.0.116

Usage

import { createBentley } from "@coopah/bentley-core";
import { bentleyOllamaPlugin } from "@coopah/bentley-provider-ollama";

const bentley = createBentley({
  plugins: [
    bentleyOllamaPlugin({ baseUrl: "http://localhost:11434/v1" }),
    // or: bentleyOllamaPlugin() — defaults to http://localhost:11434/v1
  ],
});

Then reference Ollama models in your shell config (e.g. llama3.2, mistral, gemma2).

API

• bentleyOllamaPlugin(options?) — BentleyPlugin that registers Ollama as an LLM provider
• createBentleyOllamaProvider(baseUrl?) — Low-level factory returning (modelId: string) => LanguageModel

Configuration

interface OllamaPluginOptions {
  baseUrl?: string; // Default: "http://localhost:11434/v1"
}

Related Packages

| Package | Role | |---------|------| | @coopah/bentley-core | Core runtime (required) | | @coopah/bentley-provider-openai | OpenAI provider | | @coopah/bentley-provider-anthropic | Anthropic provider | | @coopah/bentley-provider-copilot | GitHub Copilot provider |

License

MIT