Live Coding a Streaming ChatGPT Proxy with Swift OpenAPI from Scratch

Watch this 20-minute conference talk from try! Swift Tokyo 2025 where Si Beaumont demonstrates live coding a streaming ChatGPT proxy using Swift OpenAPI Generator from scratch. Learn how to build a complete system that includes generating a type-safe ChatGPT macOS client with URLSession OpenAPI transport, streaming LLM responses using Server Sent Events (SSE), creating a Linux proxy server with Vapor OpenAPI transport, reusing the same generated client within the proxy by switching to AsyncHTTPClient transport, and efficiently transforming responses while maintaining end-to-end streaming without buffering. This presentation is part of the three-day try! Swift Tokyo event (April 9-11, 2025) where developers worldwide gather to share Swift development tips, tricks, and case studies.