Hey everyone, I’m working on a workflow where Bubble sends data to Make.com to generate an AI-based report. The processing time on Make can be long, so I’m exploring the best pattern for handling the response. One suggestion is the usual POST → POST pattern (Bubble POSTs data to Make, Make POSTs the final result back to Bubble). This is simple, but it requires exposing a Bubble API endpoint publicly—technically safe with authentication, but not always preferred.
Another option is POST → GET, where Bubble sends the job via POST and Make stores the processed result in its internal Data Stores, which Bubble can then poll for using GET until the job is complete. This avoids exposing Bubble’s API but relies on Make-specific features and might require re-architecting later if migrating to n8n (which doesn’t have Data Stores).
Has anyone implemented either pattern for long-running processes? Which approach has worked best for you, and why?