> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tydli.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Welcome to Tydli

> Deploy your API docs as AI tools — instantly

# 🚀 Tydli — Turn Any API Into an MCP Server Instantly

Tydli makes your APIs **AI-accessible**.\
With one click, you can turn *any* API into a secure, Model Context Protocol (MCP) server that works with AI clients like **Claude**, **ChatGPT**, **Lovable.dev**, and other MCP-compatible tools.

***

## 🎯 What Tydli Does

Tydli converts your API documentation (OpenAPI, Swagger, JSON, or even PDFs/DOCX) into a fully operational **MCP server**.\
This allows AI systems to safely interact with your backend using standardized, secure protocols.

***

## ⭐ Key Features

* ✅ **MCP 2025-03-26 compliant**
* ✅ **Supports all major AI clients** (Claude, ChatGPT, Lovable.dev, and more)
* ✅ **OAuth 2.1 with PKCE** authentication
* ✅ **Secure credential encryption**
* ✅ **Email verification for account security**
* ✅ **AI-powered document extraction** (convert PDFs & DOCX files into API specs)
* ✅ **Real-time Analytics Dashboard** — monitor usage, performance, and errors
* ✅ **One-click deployment** — live MCP server in under 60 seconds
* ✅ **Automatic metadata discovery** via `/.well-known/mcp-server`

***

## ⚙️ How It Works

1. **Upload Your API**
   * OpenAPI/Swagger
   * JSON specs
   * URL import
   * Or upload a PDF/DOCX and let AI extract the API structure

2. **Validation**\
   Tydli automatically detects endpoints, request bodies, parameters, and authentication schemes.

3. **Configure Authentication**\
   Add API keys, OAuth credentials, or custom headers as needed.

4. **Deploy**\
   Your MCP server goes live instantly.

5. **Connect to AI Clients**\
   Use OAuth or JWT to connect applications like Claude, ChatGPT, or Lovable.dev.

***

## 🔐 Authentication Methods

### **OAuth 2.1 with PKCE** (Recommended)

This is the preferred authentication flow for most users.

Typical setup:

1. Go to the **Connectors** section inside your AI client
2. Select **Add Custom Connector**
3. Enter the MCP server URL and OAuth details provided in your Tydli dashboard
4. Approve the request in your browser
5. Done — your AI can now call your API securely

**Note:** PKCE does *not* require a client secret and is safe for public clients.

***

### **JWT Bearer Tokens** (Alternative)

If you'd like to integrate programmatically, you can generate a JWT token in your Tydli dashboard and use it directly.\
Your AI client or script will include the token in each request, and the MCP server will validate it automatically.

***

## 🚀 Quick Start for New Users

1. Sign up at **[https://tydli.io](https://tydli.io)**
2. Verify your email
3. Upload or generate your API specification
4. Click **Deploy MCP Server**
5. Connect your AI client using OAuth or JWT

Your API is now accessible to AI tools.

***

## 📚 Documentation

**Quick Links:**

* [User Guide](/user/user_guide) — Complete usage instructions
* [Security Documentation](/user/security) — Credential encryption & security architecture
* [OAuth Integration Guide](/user/oauth_integration_guide) — OAuth 2.1 setup
* [MCP Compliance](/api/mcp_compliance) — MCP protocol details
* [Rate Limits](/api/rate_limits) — API rate limiting
* [Troubleshooting](/user/troubleshooting) — Common issues

***

## 🟢 Production Status

Tydli is **production-ready** and actively maintained.

Highlights:

* ✔ Email verification
* ✔ OAuth 2.1 PKCE authentication
* ✔ JWT token support
* ✔ Document-to-API extraction
* ✔ Credential encryption
* ✔ Rate limiting
* ✔ Analytics Dashboard
* ✔ Deployment monitoring
* ✔ Mobile-responsive interface

***

# ✨ Tydli — Making APIs AI-Accessible

Turn any API into a secure MCP server in seconds.

**Website:** [https://tydli.io](https://tydli.io)