Cursor - 老张API

Cursor is an AI-native code editor built on VS Code, integrating powerful AI capabilities to provide developers with an unprecedented programming experience. Through Laozhang API, you can use various advanced large language models in Cursor.

Quick Start

1. Download and Install

Visit Cursor Official Website to download and install the version suitable for your operating system:

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

2. Configure Laozhang API

Open Settings

Click “Cursor” in the menu bar > “Settings” (or use shortcut Cmd/Ctrl + ,)

Find Models Configuration

In the left sidebar, select “Models” option

Configure OpenAI Compatible API

Scroll to the “OpenAI API Key” section:

API Provider: Select “OpenAI Compatible”
Base URL: Enter https://api2.laozhang.ai/v1
API Key: Enter your Laozhang API key
Default Model: Choose a model (e.g., gpt-5.5)

Save Configuration

Click “Save” button to complete configuration

Obtaining API Key

Visit Laozhang API Console
Click “Create New Token” button
Choose a suitable billing method
Copy the generated API Key

Recommended Model Selection

Cursor supports multiple large language models, choose appropriate models based on different programming needs:

Programming Scenario	Recommended Model	Cost	Performance
Daily Code Completion	`gpt-4.1-mini`	Low	Fast response
Complex Function Development	`gpt-5.5`	Medium	Strong reasoning ability
Large Project Refactoring	`claude-sonnet-4-6`	Medium	Long context support
Multi-language Programming	`gemini-3.1-pro-preview`	Medium	Broad language support
Code Review and Optimization	`gpt-5.5`	Medium	Deep code understanding

Model Selection Strategy

Cost-sensitive users: Recommend gpt-4.1-mini for daily coding
Quality prioritizers: Recommend gpt-5.5 or claude-sonnet-4-6
Large codebase: Recommend claude-sonnet-4-6
Multimodal needs: Recommend gemini-3.1-pro-preview, supports image understanding

Core Features

1. AI Code Completion (Tab)

Cursor provides intelligent code completion triggered by the Tab key: Features:

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

Usage Example:

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

2. AI Chat (Cmd/Ctrl + K)

Chat with AI in the editor, get instant help: Use Cases:

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

Usage Example:

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

3. AI Editing (Cmd/Ctrl + L)

Select code, trigger AI editing mode: Features:

Code refactoring
Style adjustment
Add comments
Type annotation
Error fixing

Usage Example:

# 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

Cursor can index your entire codebase, providing:

Global code search
Intelligent jump to definition
Refactoring suggestions
Dependency analysis

5. Multi-file Editing

Operate on multiple files simultaneously:

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

Advanced Features

@Symbols Citation

Use @ symbol in chat mode to cite code elements:

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

Example:

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

Terminal Integration

Cursor integrates terminal, you can:

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

Git Integration

Deeply integrated with Git:

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

Configuration Best Practices

Basic Configuration

{
  "cursor.ai.baseUrl": "https://api2.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://api2.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

Configure different models for different scenarios:

{
  "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

Use AI Chat to discuss project architecture
Let AI generate project scaffolding
Use code completion to write core logic
Use AI editing to optimize code structure
Let AI generate unit tests

Scenario 2: Code Refactoring

Select code needing refactoring
Use Cmd+L to trigger editing mode
Describe refactoring requirements
Review AI suggestions
Iterative optimization

Scenario 3: Bug Fixing

Copy error message to AI chat
AI analyzes error cause
AI provides repair suggestions
Test repair effectiveness
Let AI generate test cases to prevent regression

Scenario 4: Learning New Technologies

Ask AI how to use new library
Request sample code
Let AI explain key concepts
Use code completion to practice
Request best practice suggestions

Troubleshooting

Connection Issues

Problem: Unable to connect to API Solutions:

Check if Base URL is correct: https://api2.laozhang.ai/v1
Verify API Key validity (visit console)
Check network connection
Confirm firewall settings

No Response

Problem: AI doesn’t respond Solutions:

Check account balance
Confirm model availability
Try switching models
Check if request exceeds rate limit

Inaccurate Suggestions

Problem: Code suggestions don’t meet expectations Solutions:

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

Response Too Slow

Problem: AI response speed is slow Solutions:

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

Best Practices

1. Effective Communication

Communicate clearly and specifically with AI:

❌ 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

Make full use of Cursor’s context awareness:

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

3. Iterative Improvement

Don’t expect perfect results at once:

Get initial suggestions
Test and evaluate
Provide feedback
Iterative optimization

4. Security Awareness

Protect sensitive information:

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

5. Cost Control

Reasonably control usage costs:

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

Shortcut Reference

Shortcut	Function
`Tab`	Trigger code completion
`Cmd/Ctrl + K`	Open AI chat
`Cmd/Ctrl + L`	Trigger AI editing
`Cmd/Ctrl + Shift + P`	Open command panel
`Cmd/Ctrl + P`	Quick file open
`Cmd/Ctrl + ,`	Open settings

Further Learning

Official Resources

Learning Resources

Community Support

Need more help? Please visit Laozhang API Official Website or contact our support team.

​Quick Start

​1. Download and Install

​2. Configure Laozhang API

​Recommended Model Selection

​Core Features

​1. AI Code Completion (Tab)

​2. AI Chat (Cmd/Ctrl + K)

​3. AI Editing (Cmd/Ctrl + L)

​4. Codebase Indexing

​5. Multi-file Editing

​Advanced Features

​@Symbols Citation

​Terminal Integration

​Git Integration

​Configuration Best Practices

​Basic Configuration

​Advanced Configuration

​Multi-Model Configuration

​Usage Scenarios

​Scenario 1: New Project Development

​Scenario 2: Code Refactoring

​Scenario 3: Bug Fixing

​Scenario 4: Learning New Technologies

​Troubleshooting

​Connection Issues

​No Response

​Inaccurate Suggestions

​Response Too Slow

​Best Practices

​1. Effective Communication

​2. Utilize Context

​3. Iterative Improvement

​4. Security Awareness

​5. Cost Control

​Shortcut Reference

​Further Learning

​Official Resources

​Learning Resources

​Community Support

Quick Start

1. Download and Install

2. Configure Laozhang API

Recommended Model Selection

Core Features

1. AI Code Completion (Tab)

2. AI Chat (Cmd/Ctrl + K)

3. AI Editing (Cmd/Ctrl + L)

4. Codebase Indexing

5. Multi-file Editing

Advanced Features

@Symbols Citation

Terminal Integration

Git Integration

Configuration Best Practices

Basic Configuration

Advanced Configuration

Multi-Model Configuration

Usage Scenarios

Scenario 1: New Project Development

Scenario 2: Code Refactoring

Scenario 3: Bug Fixing

Scenario 4: Learning New Technologies

Troubleshooting

Connection Issues

No Response

Inaccurate Suggestions

Response Too Slow

Best Practices

1. Effective Communication

2. Utilize Context

3. Iterative Improvement

4. Security Awareness

5. Cost Control

Shortcut Reference

Further Learning

Official Resources

Learning Resources

Community Support