Elysia LLMs TXT

An Elysia plugin that generates llms.txt documentation from OpenAPI/Swagger specifications. Transform your API specs into LLM-friendly documentation with automatic endpoint generation.

Why Elysia LLMs TXT?

Elysia LLMs TXT bridges the gap between your API documentation and AI consumption, making it perfect for:

• LLM Integration: Generate AI-friendly documentation that large language models can easily parse
• API Documentation: Convert OpenAPI/Swagger specs into readable markdown format
• Developer Experience: Provide /llms.txt endpoints for instant API documentation access
• AI Training: Create structured documentation suitable for training AI models on your API
• Documentation Automation: Automatically generate and serve API documentation without manual maintenance

Key Features

• OpenAPI/Swagger Support: Works with both file-based and URL-based OpenAPI specifications
• JSON & YAML Support: Automatically detects and parses both JSON and YAML formats
• LLM-Optimized Output: Generates markdown documentation specifically formatted for AI consumption
• Elysia Native: Built specifically for Elysia with TypeScript support
• Flexible Configuration: Customizable headers, footers, and content types
• Standard Endpoints: Provides /llms.txt and /llms-full.txt endpoints automatically
• Built-in Caching: Optional LRU cache with configurable TTL and max size

Installation

npm install @opuu/elysia-llms-txt

pnpm add @opuu/elysia-llms-txt

yarn add @opuu/elysia-llms-txt

bun add @opuu/elysia-llms-txt

Quick Start

1. Basic Usage

import { Elysia } from "elysia";
import { llms } from "@opuu/elysia-llms-txt";

const app = new Elysia()
  .use(
    llms({
      source: {
        type: "url",
        url: "/swagger/json",
      },
    })
  )
  .listen(3000);

// Visit: http://localhost:3000/llms.txt

2. File-based Configuration

import { Elysia } from "elysia";
import { llms } from "@opuu/elysia-llms-txt";

const app = new Elysia()
  .use(
    llms({
      source: {
        type: "file",
        file: "./openapi.json",
      },
      header: "# My API Documentation",
      footer: "Generated for LLM consumption",
    })
  )
  .listen(3000);

3. Advanced Configuration

import { Elysia } from "elysia";
import { llms } from "@opuu/elysia-llms-txt";

const app = new Elysia()
  .use(
    llms({
      source: {
        type: "url",
        url: "https://petstore.swagger.io/v2/swagger.json",
      },
      header: `# Petstore API Documentation
    
This documentation is optimized for Large Language Models.`,
      footer: `
---
Last updated: ${new Date().toISOString()}
Generated by @opuu/elysia-llms-txt`,
      contentType: "text/markdown",
    })
  )
  .listen(3000);

API Reference

Configuration Options

The plugin accepts an LLMsOptions object with the following properties:

Option	Type	Description	Default
source	object	Source configuration for OpenAPI spec	Required
source.type	"file" | "url"	Whether to load from file or URL	Required
source.file	string	File path (when type is "file")	Optional
source.url	string	URL path (when type is "url")	Optional
header	string	Optional header content	Optional
footer	string	Optional footer content	Optional
contentType	"text/markdown" | "text/plain"	Response content type	"text/markdown"
cache	object	Cache configuration	Optional
cache.enabled	boolean	Enable response caching	false
cache.ttl	number	Cache TTL in milliseconds	60000
cache.maxSize	number	Max cache entries (LRU eviction)	100

YAML Support

The plugin automatically detects and parses YAML OpenAPI specs. To enable YAML support, install js-yaml:

npm install js-yaml

YAML files are detected by:

1. File extension (.yaml or .yml)
2. Content analysis (if not starting with { or [)

Generated Endpoints

The plugin automatically adds these endpoints to your Elysia application:

• GET /llms.txt - Returns the generated LLM documentation
• GET /llms-full.txt - Redirects to /llms.txt (for compatibility)

Requirements

• Node.js >= 18.0.0
• Bun >= 1.0.0
• Elysia >= 1.0.0

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.