Cursor

Documentation Index

Fetch the complete documentation index at: https://docs.laozhang.ai/llms.txt

Use this file to discover all available pages before exploring further.

​

Quick Start

​

1. Download and Install

• Windows: Windows 10 or higher
• macOS: macOS 10.15 or higher
• Linux: Ubuntu 18.04 or higher

​

2. Configure Laozhang API

​

Recommended Model Selection

​

Core Features

​

1. AI Code Completion (Tab)

• Context-aware smart completion
• Multi-line code generation
• Automatic import statement generation
• Intelligent comment writing

# Type a comment describing functionality, press Tab
def calculate_fibonacci(n):
    # Cursor automatically generates complete function code

​

2. AI Chat (Cmd/Ctrl + K)

• Ask code questions
• Request refactoring suggestions
• Generate test cases
• Explain complex algorithms
• Debug error messages

You: How to optimize this function's performance?
AI: Analyzes code and provides specific optimization suggestions

​

3. AI Editing (Cmd/Ctrl + L)

• Code refactoring
• Style adjustment
• Add comments
• Type annotation
• Error fixing

# Select this function, use Cmd+L
def calc(a, b):
    return a + b

# AI can:
# 1. Add type hints
# 2. Write docstrings
# 3. Optimize code structure

​

4. Codebase Indexing

• Global code search
• Intelligent jump to definition
• Refactoring suggestions
• Dependency analysis

​

5. Multi-file Editing

• Batch code refactoring
• Consistent style modification
• Cross-file function extraction
• Unified dependency updates

​

Advanced Features

​

@Symbols Citation

@filename - Cite specific file
@function - Cite specific function  
@class - Cite specific class
@folder - Cite entire directory

You: @utils.py How should this file be refactored?
AI: Analyzes the file and provides refactoring suggestions

​

Terminal Integration

• Let AI generate terminal commands
• Explain command outputs
• Debug errors
• Optimize scripts

​

Git Integration

• AI-generated commit messages
• Code review assistance
• Conflict resolution suggestions
• Branch management advice

​

Configuration Best Practices

​

Basic Configuration

{
  "cursor.ai.baseUrl": "https://api.laozhang.ai/v1",
  "cursor.ai.apiKey": "Your Laozhang API key",
  "cursor.ai.model": "gpt-5.5",
  "cursor.ai.maxTokens": 2000,
  "cursor.ai.temperature": 0.7
}

​

Advanced Configuration

{
  "cursor.ai.baseUrl": "https://api.laozhang.ai/v1",
  "cursor.ai.apiKey": "Your Laozhang API key",
  "cursor.ai.model": "gpt-5.5",
  "cursor.ai.maxTokens": 4000,
  "cursor.ai.temperature": 0.7,
  "cursor.ai.enableCodeLens": true,
  "cursor.ai.enableInlineCompletion": true,
  "cursor.ai.autoSuggest": true,
  "cursor.ai.contextLines": 50
}

​

Multi-Model Configuration

{
  "cursor.ai.completionModel": "gpt-4.1-mini",  // Code completion uses lightweight model
  "cursor.ai.chatModel": "gpt-5.5",          // Chat uses powerful model
  "cursor.ai.editModel": "claude-sonnet-4-6"     // Editing uses long context model
}

​

Usage Scenarios

​

Scenario 1: New Project Development

1. Use AI Chat to discuss project architecture
2. Let AI generate project scaffolding
3. Use code completion to write core logic
4. Use AI editing to optimize code structure
5. Let AI generate unit tests

​

Scenario 2: Code Refactoring

1. Select code needing refactoring
2. Use Cmd+L to trigger editing mode
3. Describe refactoring requirements
4. Review AI suggestions
5. Iterative optimization

​

Scenario 3: Bug Fixing

1. Copy error message to AI chat
2. AI analyzes error cause
3. AI provides repair suggestions
4. Test repair effectiveness
5. Let AI generate test cases to prevent regression

​

Scenario 4: Learning New Technologies

1. Ask AI how to use new library
2. Request sample code
3. Let AI explain key concepts
4. Use code completion to practice
5. Request best practice suggestions

​

Troubleshooting

​

Connection Issues

1. Check if Base URL is correct: https://api.laozhang.ai/v1
2. Verify API Key validity (visit console)
3. Check network connection
4. Confirm firewall settings

​

No Response

1. Check account balance
2. Confirm model availability
3. Try switching models
4. Check if request exceeds rate limit

​

Inaccurate Suggestions

1. Provide more context information
2. Refine requirement descriptions
3. Try different models
4. Adjust temperature parameter (lower = more deterministic, higher = more creative)

​

Response Too Slow

1. Switch to a faster model (e.g., gpt-4.1-mini)
2. Reduce maxTokens setting
3. Simplify query content
4. Check network speed

​

Best Practices

​

1. Effective Communication

❌ Bad example: "Optimize this code"
✅ Good example: "Optimize this sorting function's time complexity, preferably to O(n log n)"

❌ Bad example: "Write a function"
✅ Good example: "Write a Python function that receives a list of User objects, filters active users, sorts by registration time, and returns top 10"

​

2. Utilize Context

• Use @ symbol to cite relevant files
• Provide necessary background information
• Keep editor open with related code

​

3. Iterative Improvement

1. Get initial suggestions
2. Test and evaluate
3. Provide feedback
4. Iterative optimization

​

4. Security Awareness

• Don’t paste actual API keys in chat
• Don’t share production database connection strings
• Don’t upload confidential business logic

​

5. Cost Control

• Use lighter models for simple tasks
• Avoid unnecessary repeated requests
• Set reasonable maxTokens limits
• Regularly review usage

​

Shortcut Reference

​

Further Learning

​

Official Resources

• Cursor Official Documentation
• Cursor Community Forum
• Cursor YouTube Channel

​

Learning Resources

• Cursor Quick Start Video
• Cursor Usage Tips
• Cursor Best Practices

​

Community Support

• Cursor Discord Community
• Cursor GitHub Discussions