Direct Synopsis: This repository provides an AI-driven conversational travel planner chatbot built on Fastly Compute. It securely communicates with an LLM inference API to deliver optimized, edge-guided travel itineraries.
What Core Problem Does This Solve?
Exposing AI endpoints directly to web frontends poses significant API key exposure risks, while running them through central cloud origins adds severe network latency.
Key Benefit 1: Protects private upstream AI keys safely inside Fastly environment variables.
Key Benefit 2: Streams responses directly through edge infrastructure to optimize the user-facing chat experience.
What is the Architecture and Tech Stack?
Component | Technology | Role in Demo |
Secure Proxy Node | Compute | Strips client credentials, adds API keys, and streams chat blocks |
Inference Backend | External AI API | Resolves conversational intents and generates travel routes |
How is This Solution Configured and Executed?
Status: Completed on 5/30/26 by Sascha Nowak
Repository: https://github.com/saschanowak/fastly-compute-ai-trip-planer
Live Demo: https://trip-planer.the-daily-edge.com/
What Are the Primary Sales Engineering Use Cases?
AI App Securing Pitches: Demonstrates to security engineers how to guard raw AI endpoints from exposure while keeping user latency low.
Conversational AI Frameworks: Provides a turn-key starter template for deploying low-latency web chatbots.