elizaos-plugin-ollama 2.0.0a4

elizaOS Ollama Plugin - Local LLM client for text and object generation

elizaOS Plugin Ollama (Python)

Ollama plugin for elizaOS - Local LLM client for text and object generation.

Installation

pip install elizaos-plugin-ollama

Usage

from elizaos_plugin_ollama import OllamaClient, OllamaConfig

# Create a client with default configuration
config = OllamaConfig.from_env()
client = OllamaClient(config)

# Generate text
response = await client.generate_text_large("What is the meaning of life?")
print(response.text)

# Generate embeddings
embedding = await client.generate_embedding("Hello, world!")
print(f"Embedding dimension: {len(embedding.embedding)}")

Configuration

The client can be configured via environment variables:

Variable	Description	Default
OLLAMA_API_ENDPOINT	Base URL for the Ollama API	http://localhost:11434
OLLAMA_SMALL_MODEL	Model for small text generation	gemma3:latest
OLLAMA_LARGE_MODEL	Model for large text generation	gemma3:latest
OLLAMA_EMBEDDING_MODEL	Model for embeddings	nomic-embed-text:latest
OLLAMA_TIMEOUT_SECONDS	Request timeout	300

Features

• Text Generation: Generate text using small or large models
• Object Generation: Generate structured JSON objects
• Text Embeddings: Generate embeddings for semantic search
• Model Management: Automatic model downloading if not available locally

Requirements

• Python 3.11+
• Running Ollama server (https://ollama.ai)

License

MIT