Better parallel processing for AI calls in Frappe background jobs?

Frappe Framework App Development
,
1

Hi everyone,

My Frappe app receives frequent triggers from another system and each job calls an AI API.

I’m using frappe.enqueue(), but I noticed jobs in the queue run sequentially per worker.
To increase throughput I need to increase workers, which feels more like scaling processes than true async handling.

Is there a better pattern in Frappe for this use case?
Or is the recommended approach to offload this to an async service (for example alongside FastAPI)?

Thanks!