This guide will help you get up and running with Nia AI quickly, from authentication to asking your first questions about your codebase.

Getting Started

1

Sign Up

  1. Visit Nia AI

  2. Click “Sign In” or “Sign Up”

  3. Use your GitHub account or email for authentication

Using GitHub authentication gives you immediate access to your repositories.

2

Create Your First Project

After logging in:

  1. Click “New Project” button

  2. Enter a project name (e.g., “My Backend API”)

  3. Select the repository you want to analyze

You can connect multiple repositories to different projects to keep your codebases organized.

3

Index Your Repository

# You can specify:
- Main branch (default)
- Specific branch
- Commit hash

Initial indexing might take a few minutes depending on repository size.

Asking Questions

Effective Questions

Do Ask About:

  • Code architecture

  • Implementation details

  • Function purposes

  • Dependencies

  • Best practices

Less Effective Questions

Avoid Asking:

  • Runtime errors

  • System-specific issues

  • Questions without context

  • Non-code related queries

Question Examples

Best Practices

Project Organization

1

Separate Projects

Create different projects for:

  • Frontend codebase

  • Backend services

  • Infrastructure code

2

Regular Indexing

Re-index when you:

  • Make major changes

  • Switch branches

  • Pull updates

3

Chat Organization

  • Use descriptive chat titles

  • Keep related questions in same chat

  • Create new chats for different topics

Getting Better Answers

Be Specific

  • Reference specific files or functions

  • Provide context for your questions

  • Mention related components

Use Features

  • Follow up on previous answers

  • Use code snippets when relevant

  • Leverage chat history

Advanced Features

Choose between different AI models:

  • Claude 3 Sonnet (Default)

  • GPT-4

  • Gemini Pro

  • Deepseek V3 & R1

Each model has different strengths for code analysis.

Tips for Success

Keep Context

Stay within the same chat for related questions to maintain context

Be Patient

Allow time for thorough analysis of complex questions

Iterate

Refine questions based on previous answers

Common Questions

Example Workflow

1

Initial Setup

1. Create new project "Backend API"
2. Connect GitHub repository
3. Index main branch
2

First Questions

Ask:
"Can you explain the overall architecture of this codebase?"
"What are the main components and their responsibilities?"
3

Detailed Exploration

Follow up with:
"How is authentication implemented?"
"Show me the API endpoints structure"
4

Specific Features

Deep dive:
"Explain the error handling in the user service"
"How are database connections managed?"