Skip to main content
Zed integrates powerful AI capabilities directly into your development workflow, offering intelligent code assistance, automated editing, and conversational programming support.

​
Key AI Features

​
Agent Panel

The Agent Panel provides a conversational interface for AI-assisted development:
  • Multi-turn conversations with context awareness
  • File and directory context integration
  • Terminal command execution with AI assistance
  • Code editing with diff preview and review
  • Thread history for organizing conversations
  • Multiple AI agents including native Zed Agent and custom external agents

​
Inline Assistant

The Inline Assistant brings AI directly into your editor:
  • Code generation from natural language prompts
  • Code transformation and refactoring
  • Multi-selection support for batch operations
  • Streaming completions with real-time preview
  • Alternative suggestions from multiple models
  • Undo/redo support for AI-generated changes

​
Terminal Assistant

AI-powered terminal command assistance:
  • Command generation from natural language
  • Error explanation and debugging help
  • Command history integration
  • Safe command execution with confirmation prompts

​
Use Cases

​
Code Generation

Generate code from natural language descriptions: 
Select code region → Cmd+Enter → "Add error handling for this API call"
The AI will analyze your code context and generate appropriate error handling.

​
Code Review

Use the Agent Panel to review changes: 
@git:diff Review this branch diff for potential issues

​
Refactoring

Transform code structure with the Inline Assistant: 
"Convert this class to use dependency injection"
"Extract this logic into a separate function"
"Optimize this database query"

​
Documentation

Generate documentation from code: 
"Add JSDoc comments explaining this function's parameters and return value"

​
Debugging

Get help understanding and fixing issues: 
@diagnostics Explain these TypeScript errors and suggest fixes

​
Learning

Ask questions about your codebase: 
"How does the authentication flow work in this application?"
"Explain the purpose of this configuration file"

​
AI Providers

Zed supports multiple AI providers out of the box:
  • Anthropic (Claude models)
  • OpenAI (GPT models)
  • Google AI (Gemini models)
  • Zed Cloud (Powered by various providers)
  • Ollama (Local models)
  • Custom providers (OpenAI-compatible APIs)
See AI Providers for detailed setup instructions.

​
Model Selection

Configure different models for different tasks: 
{
  "agent": {
    "default_model": {
      "provider": "anthropic",
      "model": "claude-3-5-sonnet-20241022"
    },
    "inline_assistant_model": {
      "provider": "openai",
      "model": "gpt-4o"
    },
    "commit_message_model": {
      "provider": "zed.dev",
      "model": "claude-3-5-sonnet-20241022"
    }
  }
}

​
Context Awareness

Zed’s AI features understand your development context:
  • File contents and structure
  • Active selections and cursor position
  • Git repository state and history
  • Language server information (types, definitions, diagnostics)
  • Terminal output and history
  • Project structure and dependencies

​
Privacy & Security

  • API keys are stored securely in your system keychain
  • Code context is only sent when explicitly requested
  • Tool permissions can be configured to control AI actions
  • Local models (via Ollama) keep everything on your machine

​
Getting Started

  1. Enable AI features in Settings → Agent
  2. Configure a provider in Settings → Language Models
  3. Open the Agent Panel with Cmd+Shift+A (macOS) or Ctrl+Shift+A (Linux/Windows)
  4. Use Inline Assistant by selecting code and pressing Cmd+Enter

​
Next Steps