Cursor AI Chat Mastery

Chat Interface Overview

Opening Chat

• Cmd+L (Mac) / Ctrl+L (Windows): Open chat
• Cmd+Shift+L: New chat conversation
• Click chat icon in sidebar

Chat Panel Layout

┌─────────────────────────────────────────┐
│ Model Selector │ Context │ Settings    │
├─────────────────────────────────────────┤
│                                         │
│  Conversation History                   │
│                                         │
├─────────────────────────────────────────┤
│  Input Box                              │
│  [@files] [images] [code blocks]        │
└─────────────────────────────────────────┘

Effective Chat Patterns

Code Explanation

Select code → Cmd+L →
"Explain what this code does step by step"
"What's the time complexity of this function?"
"Are there any bugs in this code?"

Code Generation

"Write a function that [description]"
"Create a class that [description]"
"Implement [algorithm/pattern] in [language]"

Refactoring Requests

Select code → Cmd+L →
"Refactor this to use async/await"
"Make this more readable"
"Extract this into smaller functions"
"Convert to TypeScript with proper types"

Debugging

"Why is this code throwing [error]?"
"Debug this function - it returns [wrong result]"
"Find the bug: expected [X] but got [Y]"

Context Management

Adding Context with @-mentions

@filename.py - Include specific file
@folder/ - Include folder contents
@codebase - Search entire codebase
@docs - Reference documentation
@web - Search the web
@git - Git history context

Context Best Practices

Good: "@auth.py How does user authentication work?"
Good: "Looking at @utils/helpers.js, how can I optimize?"

Bad: "How does my code work?" (no context)
Bad: "Fix the bug" (too vague)

Clearing Context

• Start new chat for fresh context
• Click "Clear" to reset conversation
• Context persists within conversation

Chat Commands

Special Commands

/edit - Enter edit mode
/explain - Explain selected code
/fix - Fix errors in selection
/test - Generate tests
/docs - Generate documentation
/commit - Generate commit message

Model Selection

Available Models

GPT-4 Turbo - Best quality, slower
GPT-4 - High quality
GPT-3.5 Turbo - Fast, good for simple tasks
Claude 3.5 Sonnet - Great for explanations
Claude 3 Opus - Highest quality Claude

When to Use Each

Complex architecture → GPT-4 or Claude Opus
Quick fixes → GPT-3.5 Turbo
Explanations → Claude Sonnet
Code review → GPT-4 Turbo

Advanced Chat Techniques

Multi-Turn Conversations

Turn 1: "Create a user authentication system"
Turn 2: "Add password reset functionality"
Turn 3: "Include rate limiting"
Turn 4: "Write tests for all of this"

Iterative Refinement

"Make it more efficient"
"Add error handling"
"Use a different approach"
"Simplify this"

Code Application

After AI generates code:
- Click "Apply" to insert at cursor
- Click "Copy" to clipboard
- Click "Insert" to add to file
- Use Composer for multi-file changes

Chat History

Accessing History

• Scroll up in chat panel
• Chat history persists per session
• Export conversations if needed

Managing History

- New Chat (Cmd+Shift+L) for fresh start
- Keep related questions in same chat
- Separate unrelated topics

Tips for Better Responses

Be Specific

Vague: "Make this better"
Specific: "Optimize this function for memory usage by avoiding array copies"

Provide Context

Weak: "Fix the error"
Strong: "Fix this TypeError: cannot read property 'map' of undefined
        when calling processUsers with empty response"

Break Down Complex Requests

Instead of: "Build a complete auth system"
Do:
1. "Design the database schema for users"
2. "Create the registration endpoint"
3. "Implement login with JWT"
4. "Add password reset flow"