[AI in Practice] Deploying Song Lingo to Cloud Run: Making a Private Lyrics Website Just for Me

Developer Evan Lin shares his experience deploying 'Song Lingo,' a personal AI-powered tool that transcribes YouTube lyrics and adds grammar notes, to Google Cloud Run. The project focuses on ensuring strict private access, allowing the developer to use the tool on mobile devices without exposing it to the public. Lin emphasizes a multi-layered security approach, combining Identity-Aware Proxy (IAP) with in-app verification to prevent unauthorized access. Key technical decisions included mounting a private Cloud Storage bucket as a local folder to avoid code rewrites and configuring the service to run on a single instance to manage costs and state effectively. The article highlights common pitfalls, such as the risks of using signed URLs for private content and the necessity of verifying security configurations through multiple layers. Ultimately, the project demonstrates how to securely host personal AI applications on cloud infrastructure while maintaining strict control over data and API usage.