This repository provides custom nodes for ComfyUI, enabling seamless integration with Ollama for Large Language Model (LLM) inference and experimentation. It targets ComfyUI users looking to incorporate LLM capabilities into their visual workflows, offering a user-friendly way to leverage powerful AI models.

How It Works

The nodes interact with a running Ollama server via the ollama Python client. Key nodes include OllamaGenerateV2 for text generation with system prompts and context management, OllamaConnectivityV2 for server connection, and OllamaOptionsV2 for fine-grained control over API parameters. A OllamaVision node is also available for querying images with vision-capable models. The V2 nodes offer a more modular approach to chained LLM interactions.

Quick Start & Requirements

• Installation: Use ComfyUI Manager to search and install, or git clone into custom_nodes/comfyui-ollama.
• Prerequisites: A running Ollama server accessible from the ComfyUI host. Ollama installation instructions are provided for macOS, Windows, Linux, and Docker.
• Dependencies: pip install -r requirements.txt.
• Docs: Usage examples are available in the README.

Highlighted Details

• V2 Nodes: Offer modular and controllable chained LLM flows.
• Context Preservation: OllamaGenerateAdvance and OllamaGenerateV2 support context for chaining.
• Image Input: OllamaGenerateV2 supports image inputs for multimodal workflows.
• API Parameter Control: OllamaOptionsV2 provides full control over Ollama API parameters.

Maintenance & Community

• Contributors: stavsap is the primary contributor.
• Community: No specific community links (Discord/Slack) are provided in the README.

Licensing & Compatibility

• License: Not explicitly stated in the README.
• Compatibility: Designed for ComfyUI.

Limitations & Caveats

The README does not specify the license, which may impact commercial use. It also lacks explicit details on tested ComfyUI versions or potential compatibility issues with other custom nodes.