SwarmUI_MCP_Agent

- Exposes SwarmUI image generation as MCP tools
- Integrates a local Ollama agent (dolphin-mistral)
- Provides a Streamlit UI for prompt entry
- Automatically fetches a new session and generates images
- Generated images saved in SwarmUI's Output directory

With the SwarmUIMCPServer running and SwarmUI itself running, from the /SwarmUI_MCP_Agent directory, run: python run.py

Run setup.sh

npm install
pip install -r requirements.txt

A generate_image StructuredTool is declared in tools.py that POSTS to the /generate-image endpoint of the Express server
started by the SwarmUIServer. The description of the StructuredTool and its docstring are used by the llm
to know about the tool. Theoretically, Pydantic models should assert tool input, but the current setup doesn't work for multiple
inputs yet.