Metadata-Version: 2.4
Name: langgraph-mini-dash
Version: 1.0.1
Summary: A visual debugging dashboard and UI for LangGraph multi-agent workflows
Author-email: wiffy <melnikarthur280912@gmail.com>
License: MIT
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.9
Description-Content-Type: text/markdown
Requires-Dist: fastapi
Requires-Dist: uvicorn
Requires-Dist: langgraph
Requires-Dist: pydantic

# LangGraph Mini Dash

**A lightweight real-time visualization and debugging dashboard for LangGraph multi-agent workflows.**

<p align="center">
  <img src="https://raw.githubusercontent.com/arturbuilds/langgraph-mini-dash-assets/refs/heads/main/dashboard.png" alt="LangGraph Mini Dash Dashboard" width="100%">
</p>

LangGraph Mini Dash is a lightweight developer tool for visualizing, debugging, and monitoring LangGraph workflows through an interactive local web interface.

Monitor node execution, track performance, and understand how your AI agents interact in real time.

## Features

- **Interactive Graph Visualization** - Explore your LangGraph workflows with Cytoscape.js.
- **Real-Time Execution Monitoring** - Track node states: `idle`, `active`, and `done`.
- **Live Updates** - Receive execution updates through Server-Sent Events (SSE).
- **Execution Timing** - Measure the execution time of individual nodes.
- **Node Search** - Quickly find specific nodes in your workflow.
- **Node Filtering** - Filter nodes to navigate complex graphs.
- **Dark & Light Themes** - Switch between themes with saved preferences.
- **Local Dashboard** - Runs locally using FastAPI without external monitoring infrastructure.

## Installation

```bash
pip install langgraph-mini-dash
```

## Quick Start

Create a LangGraph workflow and pass the compiled graph to `LangGraphMiniDash`.

```python
from langgraph.graph import StateGraph, START, END
from langgraph_mini_dash.server import LangGraphMiniDash

# 1. Create a workflow
workflow = StateGraph(dict)

def my_node(state):
    return {"messages": ["Hello from the test agent!"]}

workflow.add_node("agent_node", my_node)
workflow.add_edge(START, "agent_node")
workflow.add_edge("agent_node", END)

# 2. Compile the graph
graph = workflow.compile()

# 3. Launch the dashboard
if __name__ == "__main__":
    dash = LangGraphMiniDash(graph)
    dash.run(host="127.0.0.1", port=8000)
```

Open your browser at:

**http://127.0.0.1:8000**

Your local dashboard will display the workflow graph and provide execution monitoring.

## How It Works

LangGraph Mini Dash connects your compiled LangGraph workflow to a local FastAPI dashboard.

1. **Graph Initialization** - Pass your compiled LangGraph graph to `LangGraphMiniDash`.
2. **Server Startup** - Launch the local dashboard using `dash.run()`.
3. **Execution Monitoring** - Track node execution states and timing.
4. **Live Visualization** - Observe workflow execution through the interactive graph.

## Use Cases

- Debugging multi-agent AI workflows.
- Visualizing complex LangGraph architectures.
- Monitoring agent execution in real time.
- Identifying slow workflow nodes.
- Understanding execution order and graph structure.

## License

MIT License.
