Skip to main content
Polaris includes a powerful AI assistant that can help you create files, modify code, debug issues, and answer questions about your project. The assistant appears in a sidebar and has full access to your project files.

​
Getting Started

​
Starting a Conversation

The AI assistant sidebar is located on the right side of your project workspace:
1

Open your project

Navigate to any project to see the conversation sidebar
2

Type your message

Use the input field at the bottom: “Ask Polaris anything…”
3

Send your request

Press Enter or click the send button to submit your message
Your first message automatically creates a new conversation. The AI will start processing your request immediately.

​
Conversation Features

​
Multiple Conversations

Manage multiple conversation threads for different tasks: Create a new conversation:
  • Click the + (plus) icon in the conversation header
  • Start fresh while keeping previous conversations accessible
Switch between conversations:
  • Click the History icon (clock) in the conversation header
  • Browse all past conversations for the current project
  • Select any conversation to continue where you left off
Each project has its own set of conversations, keeping your work organized by context.

​
Conversation Management

The conversation sidebar includes:
  • Conversation title - Displayed at the top (defaults to “New Conversation”)
  • History button - Access past conversations
  • New conversation button - Start a fresh thread
  • Auto-scroll - Automatically scrolls to latest messages
  • Scroll to bottom button - Appears when manually scrolled up

​
Message Actions

For assistant responses, you can:
  • Copy message - Copy the response text to clipboard (last message only)
  • View message status:
    • Thinking… - AI is processing your request
    • Request cancelled - You stopped the generation
    • Completed messages show full content

​
AI Capabilities

The AI assistant has access to powerful tools to modify your project:

​
File Operations

Ask the AI to create new files with specific content:
Create a React component called Button with primary and secondary variants
The AI will:
  • Choose the appropriate filename and extension
  • Write the file content
  • Place it in the correct directory
  • Create multiple related files if needed
Request modifications to existing files:
Add error handling to the API fetch in utils.js
The AI will:
  • Read the current file content
  • Make targeted changes
  • Preserve existing code structure
  • Update the file in place
Ask about file contents:
What does the authentication logic in auth.ts do?
The AI will:
  • Read the specified file
  • Analyze the code
  • Explain functionality in natural language
Query your project structure:
Show me all the React components in the project
The AI will:
  • List relevant files and folders
  • Organize results by directory
  • Provide file paths and types
Remove unwanted files:
Delete the old test files we don't need anymore
The AI will:
  • Identify files matching your description
  • Remove them from the project
  • Confirm what was deleted
Change file names:
Rename Button.js to PrimaryButton.js
The AI will:
  • Update the file name
  • Preserve file content and location
  • Maintain references (where possible)
Organize with new directories:
Create a components folder for all my React components
The AI will:
  • Create the folder structure
  • Place it in the appropriate location
  • Can create nested folders

​
Web Research

The AI can fetch content from web pages:
Read the documentation from https://example.com/api-docs and implement their authentication
The AI will:
  • Fetch content from the URL
  • Extract relevant information
  • Apply knowledge to your project
  • Reference multiple URLs in one request

​
Example Prompts

Here are effective ways to use the AI assistant:

​
Creating New Features

Create a contact form with name, email, and message fields. Include validation and a submit handler.

​
Debugging

Why is my API call failing? Check the fetch function in api.ts

​
Code Improvements

Refactor the user authentication code to use modern best practices

​
Learning

Explain how the state management works in App.tsx

​
Project Setup

Set up a basic Express server with CORS and environment variables
Be specific in your requests. Include file names, describe desired functionality, and mention any constraints or preferences.

​
Processing Control

​
Cancel Requests

If the AI is taking too long or you want to change direction:
1

Wait for processing

Look for the “Thinking…” indicator with spinning icon
2

Click stop button

The submit button changes to a stop icon during processing
3

Confirm cancellation

The message will show “Request cancelled” status
Cancelled requests won’t make any changes to your files. You can immediately start a new request.

​
Message Format

Assistant responses support:
  • Markdown formatting - Bold, italic, code blocks, lists
  • Code syntax highlighting - Inline and block code
  • Links - Clickable URLs
  • Structured information - Tables, nested lists

​
Best Practices

Be Specific

Include file names, exact locations, and clear requirements for better results

One Task at a Time

Break complex requests into smaller steps for more accurate execution

Provide Context

Mention related files, dependencies, or constraints the AI should consider

Review Changes

Always review AI-generated code before considering it complete

​
Limitations

The AI assistant:
  • Cannot execute code or run commands
  • Cannot access external services or APIs directly (but can read from URLs)
  • Works with file content, not runtime behavior
  • May need multiple iterations for complex changes

​
Privacy

All conversations and file operations are:
  • Scoped to the current project
  • Stored in your project database
  • Accessible only to you
  • Used to improve context for future requests in the same conversation

Build docs developers (and LLMs) love

Get started for freeTalk to us