Skip to main content

Introduction

Welcome to the Cloudflare Open Hack! Today, we’ll be diving into Cloudflare Workers AI and how to leverage Cloudflare’s ecosystem to build a serverless ETL (Extract, Transform, Load) pipeline. Our goal is to create a system that automatically processes audio files by transcribing and summarizing them using AI.

ETL Pipeline Diagram

What You’ll Learn​

  • Setting up a Cloudflare account and preparing your environment
  • Building a serverless application with Cloudflare Workers
  • Using R2 for object storage (handling audio file uploads)
  • Implementing Queues for event-driven processing
  • Storing metadata and transcripts in D1 (Cloudflare’s database)
  • Leveraging Workers AI to transcribe and summarize audio
  • Deploying and securing your application with WAF/DDoS protection

Agenda​

1. Introduction (15 min)​

Overview of Cloudflare Workers AI and how it integrates with Cloudflare’s stack.

2. Setting Up Your Cloudflare Account (10 min)​

  • Creating a Cloudflare account

3. Developing the Application - Part 1 (50 min)​

  • Configuring R2 storage
  • Setting up Workers and D1
  • Enqueueing processing tasks into Cloudflare Queues
  • Writing a Cloudflare Worker that listens for new audio files in R2

4. Break (30 min)​

Time to grab a coffee ☕ and discuss ideas!

5. Developing the Application - Part 2 (50 min)​

  • Integrating Workers AI for transcription and summarization
  • Saving transcriptions and summaries into D1
  • Handling errors and retries

6. Deploying on Cloudflare (20 min)​

  • Pushing the application live
  • Testing the pipeline

7. Security: WAF/DDoS Configuration (20 min)​

  • Adding security rules to protect your application

8. Q&A (15 min)​

Wrap-up, feedback, and next steps.