Skip to main content

​
Welcome to HAI Build

HAI Build Code Generator is an AI-powered VS Code extension that streamlines development workflows through intelligent task execution, automated code generation, and seamless integration with various LLMs.

​
Installation

Get HAI Build up and running in your VS Code environment.
1

Install from VS Code Marketplace

  1. Open Visual Studio Code
  2. Press ⇧⌘X (Mac) or Ctrl+Shift+X (Windows/Linux) to open Extensions
  3. Search for HAI Build
  4. Click Install
The extension will appear in your sidebar after installation.
2

Open the HAI Sidebar

After installation, click the HAI Build icon in the Activity Bar, or use the keyboard shortcut:
  • Mac: Cmd+'
  • Windows/Linux: Ctrl+'
This opens the HAI Build sidebar where you’ll interact with the AI assistant.
3

Configure Your AI Provider

Before you can start using HAI Build, you need to configure an LLM provider:
  1. Click the Settings icon (gear icon) in the HAI sidebar
  2. Select your preferred AI provider:
    • Anthropic (Claude)
    • OpenAI (GPT)
    • Google (Gemini)
    • OpenRouter
    • Other supported providers
  3. Enter your API key
  4. Select your preferred model
HAI Build uses a bring-your-own-key (BYOK) model, giving you full control over costs and model selection.
4

Start Your First Task

Click the + (New Task) button or press Cmd/' / Ctrl+' and describe what you want to accomplish:
Create a new React component for a user profile card
HAI will analyze your request, explore your codebase, and present a plan before making changes.

​
Understanding the Interface

The HAI Build sidebar provides several key controls:

New Task

Start a new conversation and task with HAI

HAI Tasks

View and execute AI-generated tasks from Specif AI integration

MCP Servers

Manage Model Context Protocol integrations for extended capabilities

History

Review previous tasks and conversations

Settings

Configure LLM providers, models, and extension preferences

​
Quick Actions with Keyboard Shortcuts

HAI Build provides convenient keyboard shortcuts for common actions:
ActionMacWindows/Linux
Open HAI / Jump to ChatCmd+'Ctrl+'
Add Selection to HAICmd+' (with text selected)Ctrl+' (with text selected)

​
Context Menu Actions

Right-click on selected code to access HAI Build features:
  • Add to HAI: Send selected code to the chat for discussion or modification
  • Explain with HAI: Get an explanation of what the selected code does
  • Improve with HAI: Request suggestions to improve the selected code

​
Core Workflow

Here’s how HAI Build works through a typical development task:
1

Describe Your Goal

Tell HAI what you want to achieve in natural language. Be specific about requirements and context.
2

Review the Plan

HAI analyzes your codebase and presents a clear plan outlining:
  • Files to be modified
  • Changes to be made
  • Reasoning behind decisions
3

Approve Changes

Review proposed diffs and approve changes before they’re applied. You have full visibility and control.
4

Iterate as Needed

Provide feedback, request adjustments, or ask follow-up questions. HAI maintains context throughout the conversation.

​
Best Practices

Instead of: “Make this better”Try: “Refactor this function to use async/await instead of promises and add error handling”
Help HAI understand your architecture, coding standards, and constraints:“Following our REST API patterns, create an endpoint for user authentication using JWT tokens”
Always review the proposed changes in the diff view before accepting. HAI shows you exactly what will change.
HAI creates checkpoints that allow you to easily revert changes if needed. Don’t hesitate to experiment.

​
Working with Files

HAI Build has deep codebase intelligence:
  • Automatic Discovery: HAI explores your project structure to understand your architecture
  • Context-Aware: Changes are made with awareness of your existing code patterns
  • Smart File Selection: HAI identifies relevant files based on your request
You can explicitly mention files in your requests: “Update the UserService in src/services/user.service.ts”

​
Adding Content to Chat

Multiple ways to provide context to HAI:
  1. Selected Code: Select code in the editor and press Cmd/' / Ctrl+'
  2. Context Menu: Right-click selected code → Add to HAI
  3. Terminal Output: Right-click in terminal → Add to HAI
  4. Files: Drag and drop files into the chat
  5. Screenshots: Paste images directly into the chat

​
Next Steps

Now that you’re set up, explore these powerful features:

Execute HAI Tasks

Learn how to work with AI-generated tasks from Specif AI

Code Generation

Master AI-powered code generation workflows

Git Integration

Generate commit messages and manage version control

Jupyter Notebooks

AI assistance for Jupyter notebook development
Need Help? Visit our GitHub repository or reach out at [email protected]

Build docs developers (and LLMs) love

Get started for freeTalk to us