llms-generator

llms-generator is a Node.js library designed to streamline the creation of llms.txt and llms-full.txt files. These files enable indexing of API documentation for automated and optimized use by Large Language Models (LLMs) like ChatGPT, GitHub Copilot, Anthropic's Claude, and others.

Why Use llms-generator?

The Goal

The primary goal of this library is to make your API documentation readable, structured, and AI-ready. By generating standardized files like llms.txt, you ensure that your APIs are easily indexed by LLMs, allowing developers and tools to interact with them effectively and automatically.

Why It's Important

• Enhance Developer Experience: By enabling LLMs to provide accurate, real-time responses about your APIs, you make it easier for developers to integrate with your services.
• Boost Productivity: LLMs using structured documentation can assist developers by auto-generating code, troubleshooting, and answering questions without diving into extensive manuals.
• Prepare for the Future: As AI tools become central to development workflows, having AI-optimized documentation ensures your APIs remain accessible and competitive.
• Efficiency and Cost: The llms.txt format is token-efficient, reducing processing costs while maintaining accuracy for LLM queries.

What is llms.txt?

The llms.txt standard is an emerging web convention for providing structured, concise, and optimized documentation for AI systems. It serves as a guide for LLMs to understand the structure, purpose, and usage of your APIs.

Key Features of llms.txt:

• AI Optimization: Enables seamless integration with AI developer tools.
• Structured Content: Provides information in a clear, token-efficient format.
• Interoperability: Complements standards like robots.txt and sitemap.xml.

Example llms.txt

# llms.txt - Optimized Documentation for AI Developer Tools

project_name: ExampleAPI
version: 1.0
description: An API for text analysis and sentiment scoring.
base_url: https://api.example.com
docs_url: https://docs.example.com

# API Endpoints
endpoints:
  - /v1/analyze
  - /v1/sentiment

# Authentication
auth_method: API Key
auth_details: Pass the API key in the Authorization header.

# SDKs
sdks:
  - language: Python
    sdk_url: https://github.com/example/python-sdk
  - language: JavaScript
    sdk_url: https://github.com/example/js-sdk

# Support
support_contact: [email protected]
support_url: https://support.example.com

Installation

To use llms-generator, install it via npm:

npm install llms-generator

Usage

Programmatic Usage

Generate llms.txt and llms-full.txt files programmatically:

const { generateLLMSFiles } = require('llms-generator');

// Generate files from a JSON configuration
generateLLMSFiles('./path/to/input.json', './output/directory');

Express Integration

Serve llms.txt and llms-full.txt dynamically in an Express app:

const express = require('express');
const { use } = require('llms-generator');

const app = express();
const jsonInput = {
    "project_name": "ExampleAPI",
    "version": "1.0",
    "description": "An API for text analysis and sentiment scoring.",
    "base_url": "https://api.example.com",
    "docs_url": "https://docs.example.com",
    "endpoints": [
        "/v1/analyze",
        "/v1/sentiment"
    ],
    "auth": {
        "method": "API Key",
        "details": "Pass the API key in the Authorization header."
    },
    "rate_limit": "100 requests per minute",
    "rate_limit_upgrade_url": "https://example.com/upgrade",
    "sdks": [
        {
            "language": "Python",
            "sdk_url": "https://github.com/example/python-sdk"
        },
        {
            "language": "JavaScript",
            "sdk_url": "https://github.com/example/js-sdk"
        }
    ],
    "support": {
        "contact": "[email protected]",
        "url": "https://support.example.com"
    }
};
const outputDir = './output/directory';

use(app, jsonInput, outputDir);

app.listen(3000, () => {
    console.log('Server is running at http://localhost:3000');
});

Once integrated, the following routes will be available:

• http://localhost:3000/llms.txt
• http://localhost:3000/llms-full.txt

CLI Usage

You can also run llms-generator from the command line:

npx llms-generator ./path/to/input.json ./output/directory

License

This project is open for use with the condition that credit is given to the author.

Author: Nicolas Fodor

License Terms:

You may use, modify, and distribute this project freely, provided you include the following acknowledgment in any distributed versions or derivative works:

This project uses llms-generator by Nicolas Fodor.

For commercial usage, please contact the author directly.

Contributing

Contributions are welcome! Feel free to open issues or submit pull requests.

Contact

For support or inquiries, please contact:

• Email: [email protected]