-->

⏱️ Estimated reading time: 8 min

Introduction

Devin AI has attracted significant attention among developers for its ability to carry out intern-level autonomous development tasks at $25/task. Now, however, you can achieve a comparable experience with just a $20/month Cursor or Windsurf IDE subscription, thanks to the devin.cursorrules project.

This project goes far beyond simple code completion. It brings Devin’s core capabilities - automated planning, self-evolution, web browsing, and multi-agent collaboration - into the Cursor/Windsurf environment.

What is devin.cursorrules?

devin.cursorrules is an open-source project that transforms Cursor/Windsurf IDE into an advanced AI assistant on par with Devin. Released under the MIT license, the project has earned 5.7k stars on GitHub.

Core Features

  1. Automated Planning and Self-Evolution: The AI constructs its own plans, executes them, and learns from feedback.
  2. Extended Tool Use: Web scraping, search engine queries, and LLM-based text/image analysis are invoked automatically.
  3. Multi-Agent Collaboration: A collaborative structure where o1 handles planning and Claude/GPT-4o handles execution.
  4. Cost Efficiency: Provides similar capabilities at $20/month instead of $25/task.

Detailed Feature Analysis

1. Automated Planning

Unlike conventional AI coding tools, devin.cursorrules establishes a comprehensive plan before starting any task:

  • Pre-analysis: Parses requirements and defines task steps
  • Incremental execution: Runs each step in sequence while monitoring progress
  • Dynamic adjustment: Modifies the plan and retries when problems arise during execution

2. Self-Evolution System

A system that allows the AI to learn continuously from user feedback:

  • Accumulated learning: Saves user corrections to the .cursorrules file
  • Project-specific optimization: Builds up knowledge tailored to each project’s characteristics
  • Improved iteration: Delivers increasingly accurate results over time

3. Extended Toolset

A wide range of tools comparable to Devin, invoked automatically:

  • Web scraping: Collecting web page data using Playwright
  • Search engine integration: Real-time information retrieval via DuckDuckGo
  • LLM-based analysis: Analyzing text and images to surface insights
  • Screenshot verification: UI validation through browser automation

4. Multi-Agent Collaboration

An innovative collaborative structure that improves output quality:

  • Planner: The o1 model formulates high-level plans
  • Executor: Claude/GPT-4o carries out concrete tasks
  • Cross-validation: Both agents review and refine each other’s work

Installation and Configuration Guide

The simplest and fastest setup method:

# Install cookiecutter
pip install cookiecutter

# Create a new project
cookiecutter gh:grapeot/devin.cursorrules --checkout template

Method 2: Manual Setup

When adding to an existing project:

  1. Copy required files:
    • Copy the tools folder to the project root
    • Copy the IDE-specific configuration file:
      • Windsurf: .windsurfrules, scratchpad.md
      • Cursor: .cursorrules
      • GitHub Copilot: .github/copilot-instructions.md
  2. Install dependencies: 
    pip install -r requirements.txt
  3. Set environment variables: 
    # Create .env file
cp .env.example .env
# Set API keys (optional)

Practical Usage Examples

Basic Usage

After setup, try a typical development request:

"Build a user authentication system.
Use JWT tokens and include
signup, login, and logout functionality."

The AI works through the following steps:

  1. Planning: Defines the required tech stack and implementation order
  2. Information gathering: Searches for current security best practices
  3. Code generation: Implements functionality step by step
  4. Testing: Automatically generates and runs test cases
  5. Documentation: Produces a README and API documentation

Advanced Feature Usage

"Analyze competitor websites and
identify the differentiators of our product."

For this request, the AI:

  1. Web scraping: Collects data from competitor websites
  2. Content analysis: Extracts features via LLM
  3. Comparative analysis: Identifies differences from the current product
  4. Report generation: Provides insights accompanied by visual charts

Performance Comparison and Cost Analysis

Devin vs devin.cursorrules

Feature Devin devin.cursorrules
Automated planning Yes Yes
Self-evolution Yes Yes
Web browsing Yes Yes
Multi-agent No Yes (experimental)
Monthly cost $25/task $20/month
Customizability Limited Fully open-source

Cost Efficiency

  • Devin: $250 for 10 tasks per month
  • devin.cursorrules: $20/month (unlimited tasks)
  • Savings: $230/month (92% reduction)

Use Cases and Best Practices

1. Rapid Prototype Development

"Build an MVP e-commerce platform
within 24 hours."
  • Database design
  • API server setup
  • Frontend development
  • Payment system integration
  • Deployment automation

2. Legacy Code Modernization

"Migrate this jQuery code to React."
  • Existing code analysis
  • Application of modern patterns
  • Incremental migration
  • Test code writing

3. Data Analysis and Insights

"Analyze customer data to
identify churn patterns."
  • Data preprocessing
  • Machine learning model construction
  • Visualization dashboard creation
  • Action plan proposal

Troubleshooting and Tips

Common Issues

  1. Playwright browser installation error: 
    playwright install
  2. API key configuration problems:
    • Check the .env file
    • Verify API key permissions
  3. Multi-agent conflicts:
    • Switch to single-agent mode
    • Execute the plan step by step

Performance Optimization Tips

  1. Memory usage optimization:
    • Disable unused tools
    • Use batch processing
  2. Response speed improvement:
    • Use local caching
    • Optimize parallel processing
  3. Accuracy improvement:
    • Specify requirements in detail
    • Enable step-by-step validation

Future Outlook and Roadmap

Features Under Development

  1. Improved multi-agent: Stability improvements and performance optimization
  2. Plugin system: Easy addition of custom tools
  3. Team collaboration: Multiple developers working simultaneously
  4. Cloud integration: Automated deployment to AWS, Azure, GCP

Community Contributions

  • GitHub Issues: Bug reports and feature requests
  • Pull Requests: Code contributions and improvements
  • Documentation: Usage guides and tutorial writing

Conclusion

devin.cursorrules is a project that shifts how developers approach their work. It delivers capabilities comparable to Devin’s $25/task offering at $20/month, while providing the high degree of customizability that comes with open-source software.

The multi-agent collaboration and self-evolution system are the features that most clearly set it apart from other AI coding tools. They allow developers to focus on more complex and creative work.

Start with devin.cursorrules today. Your development experience will change substantially.

Additional Resources

This post was written based on the official documentation of the devin.cursorrules project. For the latest information, check the GitHub repository.

참고