Hello everyone, I’ve got an AI voice app that I’m looking to deploy in a production-ready environment as an MVP. We’re got a basic version in AI studio, but we’re looking to move into using Vertex AI API for voice and sorting the app intro front-end / back-end infrastructure in Cloud run.
But to be honest I’m having a massive headache. I’m just wondering if anyone has any advice or ideas on just deploying this? I’ve been wracking my head for ages.
Cheers,
Probably a Fool
To deploy the AI voice app from AI Studio we can use Vertex AI and Cloud Run.
The key steps for it would be:
Switch from AI Studio to Vertex AI endpoints for text-to-speech or multimodal calls:
Generative AI on Vertex AI | Google Cloud
Backend Deployment via Cloud Run: Containerize the backend (e.g., Flask, FastAPI, Node.js) and deploy to Cloud Run:
Quickstarts | Cloud Run Documentation | Google Cloud
(*Need to set up authentication and pass your Vertex AI API key securely.)
Frontend Hosting: If we are exposing a UI, host it using Firebase Hosting or Cloud Run with a static frontend: Firebase Hosting
or reuse Cloud Run for both frontend/backend.
Security & Monitoring Setup: Use Cloud IAM to restrict API access: Identity and Access Management documentation | IAM Documentation | Google Cloud
Set up Cloud Logging/Monitoring: https://cloud.google.com/monitoring
Check quotas here: https://console.cloud.google.com/iam-admin/quotas