Cursor User Guide

I. Product Introduction

Cursor is an AI-powered code editor. By integrating Nebula Lab, you can access powerful AI assistance while writing code, supporting intelligent code generation, completion, explanation, and optimization.

II. Quick Configuration Steps

1. Open Settings

Click the gear icon ⚙️ in the top right corner and select the Models option.

2. Configure API

OpenAI API Key: Enter your Nebula Lab key (use the default token directly).

Override OpenAI Base URL: Check this box and enter https://llm.ai-nebula.com/v1.

Click Verify to confirm the configuration works.

3. Model Configuration Instructions

Important Limitation: Cursor currently does not support Agent mode, only Chat mode (you need to manually apply AI-generated code to actual files).

Suggestion for Beginners: If you rely on Agent mode for Vibe Coding, you can choose:

Purchase a Cursor official membership to use native services.

Alternatives: VS Code's RooCode or Cline extensions (both support Agent mode).

III. Recommended Model Configuration

Based on performance and cost-effectiveness, the following model configurations are recommended:

Type	Recommended Model ID	Features
Top Choice for Coding	`claude-sonnet-4-20250514`	Claude 4 Sonnet, strongest coding capability
	`gpt-4.1`	Fast speed, strong comprehensive capability
	`deepseek-v3`	Excellent in Chinese coding, high cost-performance ratio
Cost Optimization	`gpt-4.1-mini`	Lightweight but capable
	`claude-3-haiku`	Cheapest in Claude series
	`gemini-2.5-flash`	Google fast response model
Reasoning Enhancement	`o4-mini`	Top choice reasoning model for coding tasks
	`o3`	Suitable for complex reasoning and algorithmic problems

Add Custom Models

Add the following model IDs in Cursor settings:

claude-sonnet-4-20250514
gpt-4.1
deepseek-v3
o4-mini
gemini-2.5-pro

IV. Usage Mode Explanation

Chat Mode Workflow

Since Agent mode is not supported, the recommended workflow is as follows:

Generate Code via Chat:

Use Ctrl/Cmd + L to open the chat window.

Describe requirements and let AI generate code snippets.

Manually Apply Code:

Copy code from the chat window and paste it into the target file.

Or use the "Apply" button (if available).

Iterative Optimization:

Continue the dialogue to request modifications and repeat the application process.

Alternative Solutions Comparison

Tool	Agent Mode	Pros	Cons
Cursor	❌	Elegant interface, good completion experience	No Agent mode
Cline (VS Code)	✅	Full Agent features, can automatically modify files	Requires VS Code
RooCode (VS Code)	✅	Agent mode, supports multi-file editing	Newer, features still being polished
Continue (VS Code)	✅	Open source, highly customizable	Configuration is relatively complex

V. Core Functions

Intelligent Code Completion

Tab Completion: Press Tab to accept AI suggestions.

Multi-line Completion: Supports function-level code generation.

Context Awareness: Provides precise suggestions based on project structure.

AI Chat

Ctrl/Cmd + K: Open Command Palette.

Ctrl/Cmd + L: Open Sidebar Chat.

Code Explanation: Select code to directly ask AI for its meaning or optimization direction.

Code Editing

Generate Code: Describe requirements, and AI automatically generates complete code snippets.

Refactoring Suggestions: Get code structure optimization plans.

Error Fixing: AI assists in locating and fixing syntax/logic errors.

VI. Shortcuts

Shortcut	Function
`Ctrl/Cmd + K`	Open AI Command Palette
`Ctrl/Cmd + L`	Open AI Chat Sidebar
`Tab`	Accept code suggestion
`Esc`	Cancel current suggestion

VII. Usage Tips

1. Provide Clear Context

2. Optimize Prompts

Bad Example: "Fix this function"

Good Example:

"Fix the floating-point precision issue in the calculateTotal function, ensuring amounts are calculated to two decimal places"

3. Make Full Use of Chat Mode

Let AI generate complete file content.

Request detailed modification instructions.

Use for code review and refactoring suggestions.

VIII. Troubleshooting

Connection Timeout

Check if the network connection is normal.

Confirm the API URL is https://llm.ai-nebula.com/v1.

Verify the validity of the API Key.

Model Not Responding

Check if the account balance is sufficient.

Try switching to another model.

Restart the Cursor client.

Poor Quality Code Suggestions

Provide more project context (e.g., frameworks, dependency versions).

Use more specific prompts.

Try switching to a better-performing model (e.g., claude-sonnet-4-20250514).

IX. Best Practices

Project-Level Configuration

Create .cursor-settings.json in the project root directory to unify configuration rules:

{
 "model": "gpt-4.1",
 "temperature": 0.7,
 "contextFiles": ["README.md", "package.json"],
 "rules": [
  "Use TypeScript strict mode",
  "Follow ESLint standards",
  "Add appropriate comments"
 ]
}

Code Standards

Explicitly state standard requirements in prompts, for example:

Use TypeScript

Follow Airbnb standards

Add JSDoc comments

Adopt functional style

Security Awareness

Avoid including sensitive information (such as keys, tokens) in the code.

Review AI-generated code to verify logic security.

Check the security of third-party dependencies (e.g., via npm audit).

X. Integration Workflow

Git Integration

Test-Driven Development (TDD)

Write test cases first.

Let AI generate implementation code.

Run tests to verify results.

Iteratively optimize code.

Code Review

Use AI for automated review:

Please review this code, focusing on:
1. Performance issues
2. Security vulnerabilities
3. Code standards
4. Best practices

XI. Supplement on Agent Mode

If you need AI to automatically modify multiple files or perform complex refactoring tasks, the following tools are recommended:

Cline: VS Code extension, supports full Agent features, can automatically modify files.

RooCode: Emerging VS Code AI Agent extension, supports multi-file editing.

Cursor Code Editor Integration Guide

Cursor User Guide#

I. Product Introduction#

II. Quick Configuration Steps#

1. Open Settings#

2. Configure API#

3. Model Configuration Instructions#

III. Recommended Model Configuration#

Add Custom Models#

IV. Usage Mode Explanation#

Chat Mode Workflow#

Alternative Solutions Comparison#

V. Core Functions#

Intelligent Code Completion#

AI Chat#

Code Editing#

VI. Shortcuts#

VII. Usage Tips#

1. Provide Clear Context#

2. Optimize Prompts#

3. Make Full Use of Chat Mode#

VIII. Troubleshooting#

Connection Timeout#

Model Not Responding#

Poor Quality Code Suggestions#

IX. Best Practices#

Project-Level Configuration#

Code Standards#

Security Awareness#

X. Integration Workflow#

Git Integration#

Test-Driven Development (TDD)#

Code Review#

XI. Supplement on Agent Mode#