XY built an AI agent orchestration platform for healthcare using Temporal as its execution backbone. The core innovation is a YAML-based DSL interpreted by a single generic DSLWorkflow class, eliminating the need to write custom Temporal workflow code for each new healthcare automation. An LLM-powered Planner Agent translates natural language requests into validated YAML workflows. The architecture uses specialized task queues (generic, browser, GPU) with purpose-built workers that scale independently, Temporal Cloud namespaces for multi-tenant isolation, and a continue-as-new mechanism for processing thousands of records without hitting history limits. The DSL versioning approach allows safe evolution of workflow definitions while in-flight production workflows continue uninterrupted.
