🤖 Cloudflare AI Chat Assistant (Python Worker)

Note: This project was developed 100% using AI assistance as a collaborative architect. The primary goal was to build a functional, end-to-end MVP in just a few hours, focusing on simplicity and the practical integration of Cloudflare's newest features. A simple AI-powered chat application built for the Cloudflare Software Engineering Internship assignment. This project demonstrates a complete serverless flow using the Cloudflare Workers Python environment.

🚀 What it does

• Frontend: A minimalist web interface served directly from a Python Worker.
• Workflow: Processes user messages, maintains session state, and fetches AI responses.
• AI Engine: Uses Llama 3.3 on Workers AI for natural language interaction.
• Memory: Integrated with Cloudflare KV to store and retrieve chat history (contextual awareness).

🛠️ Technical Stack

• Backend: Cloudflare Workers (Python / Pyodide)
• AI Model: @cf/meta/llama-3.3-70b-instruct-fp8-fast
• Database: Workers KV (CHAT_HISTORY)

🔗 Live Demo

View Live Project

🚀 Getting Started

1. Clone the repository:

   git clone https://github.com/afonsocapabarroz/cf_ai_assistant
   cd cf_ai_assistant

2. Login to Cloudflare:

   npx wrangler login

3. Setup KV Namespace: Create the required KV namespace for chat memory:

   npx wrangler kv:namespace create "CHAT_HISTORY"

   Note: Ensure the generated id is updated in your wrangler.jsonc file.

4. Run locally:

   npx wrangler dev

5. Deploy to production:

   npx wrangler deploy