CopilotJ: A Conversational Multi-agent System for Intelligent and Efficient Bioimage Analysis

Learn more on the Project Page, follow the User Manual to get started, or watch the Video Demos to see CopilotJ in action.

About

CopilotJ turns natural-language requests into executable, verifiable bioimage analysis workflows by coordinating Fiji/ImageJ ecosystem, Python scientific libraries, and deep-learning models.

CopilotJ comprises three major components:

• Web Frontend — A browser-based chat interface through which users interact with CopilotJ. It sends requests to the multi-agent backend and receives streaming responses via RESTful APIs.
• Multi-agent Backend — Orchestrates agents and tools, integrates Python and deep-learning environments with a local database, connects to multiple LLM providers, and retrieves community knowledge sources.
• CopilotJ Bridge — A Fiji plugin that communicates with the backend via a bidirectional WebSocket API, enabling real-time exchange of system status and commands.

Visit copilotj.chat for more information and demos.

Installation & Configuration

For full installation and configuration instructions, including environment setup, API key configuration, and how to run each component, please refer to the User Manual.

License

This project is licensed under the Apache License 2.0.