Introduction

Igniter.js is the first AI-native TypeScript framework designed from the ground up for the future of software development—where humans and AI collaborate seamlessly. It provides end-to-end type safety, real-time capabilities, background jobs, and a predictable architecture that both developers and AI agents can easily understand and modify.

Built for Code Agents, Perfected for Developers

Unlike traditional frameworks that treat AI integration as an afterthought, Igniter.js was architected from day one with AI-first principles. Every design decision—from the feature-sliced architecture to the explicit type system—creates a low-entropy environment that makes it dramatically easier for Code Agents like Cursor, Claude Code, GitHub Copilot, and Gemini CLI to understand, navigate, and modify your codebase.

The framework eliminates common sources of confusion for AI agents: implicit behaviors, magic conventions, scattered configuration, and complex inheritance hierarchies. Instead, it embraces explicit over implicit, composition over inheritance, and feature-based organization that mirrors how humans naturally think about business logic.

What does this mean in practice? When you ask Claude Code to "add authentication to the users endpoint," it doesn't need to hunt through multiple files, decipher framework magic, or guess at conventions. It sees a clear, predictable structure where every feature is self-contained, every dependency is explicit, and every type is precisely defined. The AI can confidently make changes knowing exactly what will break and what won't.

But being AI-native doesn't mean sacrificing developer experience. In fact, the same principles that make code easy for AI to understand—clarity, explicitness, and strong typing—make it a joy for humans to work with too. You get exceptional IntelliSense, instant error detection, and a framework that guides you toward best practices naturally.

// Crystal-clear structure that both humans and AI understand instantly
export const users = igniter.controller({
  name: 'Users',
  description: 'Manage user accounts and profiles',
  path: '/users',
  actions: {
    getById: igniter.query({
      name: 'Get User by ID',
      description: 'Retrieve a specific user by their ID',
      path: '/:id' as const,
      handler: async ({ params }) => {
        const user = await db.users.findUnique({ where: { id: params.id } })
        return response.success({ user })
      }
    })
  }
})

// Client code is automatically type-safe—no configuration needed
const { user } = await api.users.getById.query({ id: '123' })
//     ^? User - TypeScript knows the exact type!

Why Igniter.js?

Get Started

Choose your framework to start building:

Or use the CLI for a guided setup:

npx @igniter-js/cli@latest init

pnpm @igniter-js/cli@latest init

yarn @igniter-js/cli@latest init

bunx @igniter-js/cli@latest init

AI Tooling

Igniter.js provides a complete AI development ecosystem that transforms how you build applications.

Lia - Your AI Development Partner

Lia is a specialized Code Agent trained exclusively for the Igniter.js ecosystem. Unlike generic AI assistants that struggle with framework-specific patterns, Lia is purpose-built to understand every aspect of Igniter.js—from architecture decisions to testing strategies.

When you install Lia in your project, she replaces the default behavior of Code Agents like Cursor, Claude Code, or GitHub Copilot with deep knowledge of:

• Igniter.js Architecture - Feature-based design, controller patterns, and type-safe APIs
• Development Workflows - From scaffolding to deployment, following best practices
• Testing Strategies - Unit tests, integration tests, and E2E testing patterns
• Debugging Techniques - Advanced diagnostics using MCP Server tools

What makes Lia special is her ability to not just write code, but to validate, debug, and test what she creates. She can:

• Scaffold complete features following Igniter.js conventions
• Debug type errors and suggest fixes with context-aware solutions
• Run and analyze test results to ensure quality
• Investigate code using AST parsing and semantic search
• Create GitHub issues and explore library source code
• Manage tasks with built-in project management tools

→ Learn more about Lia

MCP Server - Lia's Toolbox

The Igniter.js MCP Server is what powers Lia's advanced capabilities. It's a collection of specialized tools that transform your API into an AI-accessible workspace, enabling Lia to perform complex development tasks autonomously.

→ Explore MCP Server tools

Getting Started with Lia

Installing Lia is simple. Use the Igniter.js CLI to set up Lia in your preferred Code Agent:

npx @igniter-js/cli@latest lia --cursor

npx @igniter-js/cli@latest lia --claude

npx @igniter-js/cli@latest lia --copilot

Using Lia in Your Project

Once installed, Lia becomes your intelligent pair programmer. Here are examples of what you can ask:

Lia understands the full context of your project and delivers production-ready code that follows Igniter.js best practices.

LLMs.txt

Igniter.js website exposes multiple llms-readable documentation formats that help AI understand your API structure and how to interact with it.

Main Entry Point

/llms.txt - A comprehensive index that provides:

• Overview of the Igniter.js framework
• Site structure and content organization
• Links to all available content types
• Instructions on how to access individual pages

Visit https://igniterjs.com/llms.txt to see the complete index.

Full Content Files

For AI models that need complete documentation in a single file:

• /llms-full.txt - Complete documentation from all content types (blog, docs, learn, templates, showcase, changelog)
• /llms/blog/full.txt - All blog posts and articles
• /llms/docs/full.txt - Complete framework documentation
• /llms/learn/full.txt - All learn course chapters
• /llms/templates/full.txt - All starter templates
• /llms/showcase/full.txt - All showcase projects
• /llms/changelog/full.txt - All version updates and release notes

Individual Pages

Any page can be accessed in markdown format by appending .mdx to the URL:

• /docs/core/quick-start.mdx - Documentation page as markdown
• /blog/introducing-igniter-js.mdx - Blog post as markdown
• /learn/course/01-creating-your-project.mdx - Learn chapter as markdown

This allows AI models to fetch specific pages on-demand, reducing token usage while maintaining full context when needed.

Other Formats

• /rss - RSS feed with all content types
• /rss/[type] - RSS feed for specific content type (blog, docs, learn, etc.)
• /sitemap.xml - XML sitemap with all pages

All these formats are designed to make Igniter.js documentation easily accessible to AI models, enabling better code generation, documentation understanding, and API interaction.