I’m thrilled to share with you my latest independent project, the CrewAI Visual Editor. This tool is designed to simplify the process of creating and managing CrewAI crews through a visual, node-based interface. Whether you’re a developer or just exploring CrewAI, this editor aims to make your experience more intuitive and efficient.
Drag-and-Drop Interface: Easily create and connect nodes to build your crew structure.
Real-Time Visualization: Watch your crew setup come to life as you build it.
YAML and Python Export: Seamlessly export your configurations for further development or deployment.
Built with Modern Technologies: Utilizing React, TypeScript, and Vite for a smooth and responsive experience.
Getting Started:
Clone the Repository: Access the latest version from my GitHub repository.
Install Dependencies: Use npm or yarn to install the necessary packages.
Run the Development Server: Start building your crew with the easy-to-use interface.
How It Works:
Begin with a Node: Start by placing a Begin Node on the canvas.
Add Tasks and Agents: Drag Task and Agent Nodes onto the canvas and connect them to define your crew’s workflow.
Configure Node Properties: Fill in the details for each node to customize your crew’s behavior.
Export Your Setup: Use the export functionality to generate YAML and Python scripts for your crew.
Join the Discussion:
I would love to hear your feedback and see how you’re using the CrewAI Visual Editor in your projects. Please feel free to share your experiences, suggestions, and any questions you might have here on the forums.
Thank you for your interest in my project. I look forward to seeing the amazing crews you create with the CrewAI Visual Editor!
Since the original post, the editor has been rebuilt from the ground up. Here’s what’s new:
Redesigned Node System
Compact, color-coded nodes (amber for agents, teal for tasks, green for start) instead of bulky form-in-node cards
Dedicated Properties Panel on the right side for editing – select any node and edit all its fields without cluttering the canvas
Full CrewAI Feature Support
50+ built-in tools from the complete crewai_tools catalog (Firecrawl, Selenium, NL2SQL, vector DBs, S3, MCP adapters, and more)
Custom tool definitions – define your own Python tools or MCP Server tools in Crew Settings, and they appear in the agent tool picker alongside built-ins
LLM selection with 60+ models across OpenAI (GPT-4.1/5.x, o3/o4), Anthropic (Claude 3.7, Sonnet/Opus 4-4.6), Google Gemini (2.5/3), xAI Grok (3/4), Mistral, Groq, and OpenRouter
Crew Settings – process type (sequential/hierarchical), memory, cache, planning with planner LLM, manager LLM for hierarchical
Crew Input Variables – define {topic}, {company}, etc. that get interpolated into agent/task configs at runtime via crew.kickoff(inputs={…})
4 Ready-to-Use Templates
Research & Analysis Team
Content Production Pipeline
Marketing Campaign Team
Software Development Team
Each comes with fully configured agents (personalities, backstories, tools, LLMs) and task chains with proper {placeholder} variables.
Editor UX
Undo/Redo (Ctrl+Z/Y)
Node duplication (Ctrl+D)
Save/Load with proper modals (no more browser alerts)
Tabbed YAML + Python export with copy and download buttons
MiniMap for navigating large crews
Keyboard shortcuts throughout
Welcome screen with template quick-start when canvas is empty
Export generates production-ready code – agents.yaml, tasks.yaml, and crew.py with proper tool imports, MCP adapter setup, and a commented kickoff example showing how to pass your input variables.