QVerisFlow provides an automated, multi-agent workflow generation framework built on LangGraph. It targets engineers and power users seeking to rapidly develop, visualize, and execute complex agent-based systems. The core benefit is enabling intuitive workflow creation through natural language conversations and a visual canvas, coupled with intelligent agent generation and real-time execution monitoring.

How It Works

The framework employs a layered architecture, starting with a Definitions layer for Agent, Tool, and Workflow definitions. The Core layer provides BaseAgent and AgentGraph functionalities, enhanced by a Verification layer. The Orchestration layer features a WorkflowEngine for execution and a SmartRouter for agent selection. Crucially, the Evolution layer includes DialogueAgent and MetaAgent for conversational building and automatic Agent definition generation. The Web UI layer, built with React and ReactFlow, provides a visual interface connected via WebSockets. This multi-layered approach allows for flexible, modular development and robust execution.

Quick Start & Requirements

• Primary install: Use uv sync for dependencies or pip install -e ".[web,dev]".
• Prerequisites: Python 3.11+, Node.js/npm for the frontend.
• Configuration: Edit config/app_config.yaml to set up LLM API keys and model selections.
• Running: Start the backend with uv run python -m src.web and the frontend in a new terminal (cd web-ui && npm install && npm run dev).
• Access: Navigate to http://localhost:5173.
• Docs: Detailed documentation is available in the docs/ directory.

Highlighted Details

• Conversational Workflow Building: Create, modify, and iterate workflows using natural language prompts.
• Visual Web UI: A modern React interface featuring a drag-and-drop canvas, node editing, and real-time status display.
• Intelligent Agent Generation: The Meta Agent automatically analyzes tasks to generate optimal Agent definitions.
• Multi-Model Support: Flexible configuration for various LLMs including Qwen, DeepSeek, GPT, and GLM.
• Real-time Execution Monitoring: WebSocket enables live progress updates, with support for pausing, resuming, and intervention.
• QVeris Integration: Seamlessly connects with the QVeris unified data and tool layer.
• Three-Tier Verification: Employs rule validation, LLM judgment, and ensemble validation for robust output.

Maintenance & Community

No specific details regarding maintainers, community channels (like Discord/Slack), or roadmaps were found in the provided README.

Licensing & Compatibility

The project is licensed under the Apache License 2.0. This license is permissive and generally compatible with commercial use and linking within closed-source projects.

Limitations & Caveats

The README does not explicitly list limitations or caveats. Setup requires running both backend and frontend services concurrently. While features like the Meta Agent are highlighted, their maturity level is not specified.