Skip to content

Tutorials#

In this series of tutorials, you'll learn how to use the FastAgency framework to create interactive chatbots that can scrape the web and work with different APIs to respond to user requests.

Available Tutorials#

1. Web Scraping and Giphy API Integration#

In this tutorial, we will explore how to leverage the FastAgency framework to create a dynamic chatbot that integrates two powerful agents:

  • WebSurferAgent – A web-scraping agent capable of retrieving relevant content from webpages.
  • Giphy agent – An agent that interacts with the Giphy API to fetch GIFs based on the user’s request.

When to Use Web Scraping and Giphy API Integration?#

  • API Integration: Learn how to integrate external APIs like Giphy with FastAgency.
  • Autonomous Agents: Build and register agents that autonomously scrape the web for relevant information.
  • User-Agent Workflows: Use AutoGen workflows to manage agent interactions and user input.
  • Personalized Content: Present scraped content to users and offer personalized GIF suggestions based on that content.

Let’s dive into Web Scraping and Giphy API Integration →

2. WhatsApp API Integration and Web Scraping#

In this tutorial, we will explore how to build a chatbot using the FastAgency framework that integrates two essential agents:

  • WebSurferAgent – A web-scraping agent capable of retrieving content from websites.
  • WhatsApp agent – An agent that interacts with the Infobip WhatsApp API to send messages via WhatsApp.

When to Use WhatsApp API Integration and Web Scraping?#

  • API Integration: Learn how to integrate the Infobip WhatsApp API using OpenAPI.
  • Autonomous Agents: Build and register agents that autonomously scrape the web using the WebSurferAgent.
  • User-Agent Workflows: Manage user interactions and send scraped content via WhatsApp using AutoGen workflows.
  • Security: Handle secure API credentials using APIKeyHeader and ensure safe communication.

Let’s dive into WhatsApp API Integration and Web Scraping →

3. Agentic testing for prompt leakage security#

This tutorial introduces the Prompt Leakage Probing Framework, a tool designed to assess the security of Large Language Models (LLMs) against prompt leakage vulnerabilities. By leveraging modular and extensible components, the framework allows users to simulate various scenarios that expose sensitive information embedded in system prompts.

When to Use the Prompt Leakage Probing Framework?#

  • Evaluate Prompt Leakage Risks: Analyze LLM responses for unintentional exposure of confidential information.
  • Test Hardened Models: Compare the robustness of LLM configurations with and without security measures.
  • Simulate Scenarios: Create realistic attacks, such as Base64 encoding or targeted probing, to test LLM security.
  • Enhance Security Workflows: Use the framework’s automated detection capabilities to refine defense mechanisms for production-ready LLMs.

Let’s take a closer look at prompt leakage probing →