Ollama AI Provider Plugin

The Ollama plugin connects Ever Works to a local or remote Ollama server for self-hosted AI inference. It extends BaseAiProvider and uses the shared AiOperations layer that wraps LangChain under the hood.

Source: packages/plugins/ollama/src/ollama.plugin.ts

Overview

Property	Value
Plugin ID	ollama
Category	ai-provider
Capabilities	ai-provider
Version	1.0.0
Configuration Mode	user-required
Provider Type	ollama
Auto-enable	No
Built-in	Yes
Visibility	public

The plugin extends BaseAiProvider from @ever-works/plugin/abstract and creates an AiOperations instance from @ever-works/plugin/ai to handle all AI operations through a unified LangChain-based abstraction. Unlike cloud-hosted providers, Ollama typically does not require an API key.

Architecture

The plugin talks to Ollama through its OpenAI-compatible /v1 API endpoint. This means any model served by Ollama that exposes this endpoint works without additional configuration.

Configuration

Settings Schema

Setting	Type	Default	Scope	Description
baseUrl	string	http://localhost:11434/v1	user	Address of the Ollama instance
apiKey	string	ollama	user	Only needed for secured Ollama instances
defaultModel	string	ministral-3:8b	global	Used for all AI tasks unless a tier-specific model is set
simpleModel	string	ministral-3:8b	global	Handles tags, short descriptions, and quick classifications
mediumModel	string	ministral-3:8b	global	Handles listings, summaries, and content reformatting
complexModel	string	ministral-3:8b	global	Handles full page generation and multi-step analysis
temperature	number	0.7	hidden	Controls output randomness (0 = deterministic, 2 = creative)
maxTokens	number	4096	hidden	Maximum length of each AI-generated response

Model fields use the x-widget: model-select extension, which renders the model-select dropdown in the dashboard UI. The dropdown is populated by calling listModels() against the configured Ollama instance.

Required Fields

• baseUrl -- the Ollama server address
• defaultModel -- at least one model must be selected

Schema Extensions

{
	"defaultModel": {
		"type": "string",
		"title": "Default Model",
		"default": "ministral-3:8b",
		"x-widget": "model-select",
		"x-scope": "global"
	}
}

The x-scope: global extension means model selection applies platform-wide rather than per-user. The x-hidden: true extension on temperature and maxTokens keeps those fields out of the standard settings UI.

Model Capabilities

getCapabilities(): AiModelCapabilities {
    return {
        supportsStructuredOutput: true,
        supportsStreaming: true,
        supportsToolCalling: true,
        supportsVision: true,
        maxContextLength: 128000
    };
}

Capability	Supported
Structured output (JSON mode)	Yes
Streaming responses	Yes
Tool calling	Yes
Vision (image input)	Yes
Embeddings	Yes
Max context length	128,000 tokens

Ollama supports embeddings through models such as nomic-embed-text. The createEmbedding() method delegates to AiOperations.createEmbedding().

Tiered Model Assignment

Ever Works uses a three-tier model system to balance speed and quality during work generation:

Tier	Use Cases	Recommended Models
Simple	Tags, short descriptions, classifications	ministral-3:8b, gemma2:2b
Medium	Listings, summaries, content reformatting	ministral-3:8b, llama3.1:8b
Complex	Full page generation, multi-step analysis	llama3.1:70b, mistral-large

If a tier-specific model is not set, the defaultModel is used for all tiers.

Lifecycle

Loading

async onLoad(context: PluginContext): Promise<void> {
    await super.onLoad(context);
    this.aiOps = new AiOperations({
        apiKey: 'ollama',
        model: 'ministral-3:8b',
        baseURL: 'http://localhost:11434/v1',
        temperature: 0.7,
        maxTokens: 4096,
        providerType: 'ollama'
    });
}

On load, the plugin creates an AiOperations instance with default values. When a request arrives, resolveConfig() merges the user's saved settings on top of these defaults before executing.

Availability Check

The isAvailable() method calls AiOperations.testConnection() with the resolved configuration to verify the Ollama server is reachable and responding.

API Methods

Method	Description
createChatCompletion(options)	Single chat completion request
createStreamingChatCompletion(options)	Streaming chat completion (async generator)
createEmbedding(options)	Generate text embeddings
listModels(settings)	List available models from the Ollama instance
isAvailable(settings)	Test connection to the Ollama server
getCapabilities()	Return supported capabilities
healthCheck()	Return plugin health status

Getting Started

1. Install and run Ollama from ollama.com or connect to an existing instance.
2. Pull at least one model: ollama pull ministral-3:8b.
3. Enable the Ollama plugin in the Ever Works dashboard under Settings > Plugins.
4. Set the Ollama Server URL to your instance address (defaults to http://localhost:11434/v1).
5. Select your preferred models for each task complexity tier.
6. Ollama will be used automatically during work generation and AI conversations.

Troubleshooting

Issue	Cause	Solution
Plugin shows unavailable	Ollama server not running	Start Ollama with ollama serve
No models in dropdown	No models pulled	Run ollama pull <model-name>
Connection refused	Wrong base URL	Verify the URL includes /v1 (e.g., http://localhost:11434/v1)
Slow responses	Model too large for hardware	Use a smaller model or increase available RAM/VRAM
Embedding errors	Model does not support embeddings	Switch to an embedding model like nomic-embed-text