Connector Creator Agent

AI-powered system for creating Fastn.ai connectors with interactive chat interfaces.

What It Does

1. Web Scraping: Crawls API documentation websites to extract content using Selenium WebDriver
2. AI Endpoint Extraction: Uses gpt-5-mini to find and extract cURL commands from documentation
3. Connector Creation: Creates connector groups and endpoints in Fastn.ai with proper authentication
4. Universal Compatibility: Works with any API documentation format (REST APIs, GraphQL, etc.)
5. Interactive Chat: Conversational interface for step-by-step connector creation with session management

Setup

1. Create and activate virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

pip install -r requirements.txt

3. Install Chrome browser (required for Selenium):

   • macOS: brew install --cask google-chrome
   • Ubuntu: sudo apt-get install google-chrome-stable
   • Windows: Download from Google Chrome website

4. Copy the example environment file and configure your credentials:

cp .env.example .env

5. Edit .env file with your actual credentials:

# OpenAI API Configuration
OPENAI_API_KEY=your_openai_api_key_here

# Fastn Configuration
FASTN_ENV=qa.fastn.ai
FASTN_USERNAME=your_fastn_username
FASTN_PASSWORD=your_fastn_password
FASTN_CLIENT_ID=fastn-app
FASTN_CLIENT_SPACE_ID=your_client_space_id
FASTN_REDIRECT_URI=https://google.com

# Scraping Configuration
MAX_PAGES=10    # Maximum number of pages to scrape (default: 10)

Usage

Chat Mode (Recommended)

Interactive conversational interface with session management:

# Start new chat session
python chat_app.py

# List previous chat sessions
python chat_app.py --list

# Resume existing session
python chat_app.py --resume session_20240101_143022

Examples

Chat mode - Interactive conversation:

python chat_app.py
# Then chat: "I want to create a connector for OpenAI API at https://docs.openai.com/api"

Output

The system creates:

• Connector Groups: With proper authentication configuration
• Connector Endpoints: From extracted cURL commands
• Local Data: Scraped data and results saved in scraped_data/ directory
• Chat Sessions: Conversation history saved in conversations/ directory (chat mode only)
• Logs: Execution logs in app.log

Features

Chat Mode Features 💬

• 🗣️ Natural Conversation - Chat naturally about what connector you want to create
• 📁 Session Management - Resume previous conversations anytime
• 🔄 Persistent History - All conversations saved locally in JSON format
• 🤖 Tool Integration - AI automatically uses scraping and connector creation tools
• 📝 Session Listing - View and resume any previous chat session
• 🎯 Context Awareness - Remembers platform names and connector group IDs across conversation

Chat Workflow

When using chat mode:

1. 💬 Start Chat - Natural conversation about your needs
2. 🌐 AI Scrapes - Agent automatically scrapes documentation when you provide URL
3. 🔍 Review Endpoints - AI shows what endpoints it found
4. 🏗️ Create Group - AI suggests auth config and creates connector group
5. 🔧 Build Endpoints - AI creates each endpoint with proper cURL commands
6. 💾 Save Session - Everything saved for future reference

Technical Features

• Selenium WebDriver - Handles JavaScript-heavy documentation sites
• AI-Powered Extraction - gpt-5-mini intelligently finds API endpoints
• Multiple Auth Types - OAuth, API Key, Bearer Token, Basic Auth, Custom Input
• Smart Content Filtering - Extracts only API-relevant content from docs
• Error Handling - Robust error handling with detailed logging
• Session Persistence - Resume conversations from exactly where you left off

Requirements

• Python 3.7+
• OpenAI API key
• Chrome browser (for Selenium WebDriver)
• Internet connection
• Fastn.ai account credentials
• Access to Fastn QA environment (or production with appropriate configuration)