We are pleased to announce the following product improvements in our 2025.2 release:

• Wallaroo Inference Endpoint API Spec Generation: A new SDK method wallaroo.pipeline.Pipeline.generate_api_spec() returns the Wallaroo pipeline inference endpoint specification in yaml format under the OpenAPI 3.1.1 format. This provides AI developers an easier way to integrate applications with Wallaroo AI inference endpoints.
  • Retrieve Inference Endpoint API Spec
  • LLMs with OpenAI API Compatibility
  • Tutorials:
    • Inference Endpoint API Spec for OpenAI Compatibility Enabled vLLM Tutorial
    • Inference Endpoint API Spec Tutorial
• Wallaroo Infrastructure Usage Metrics: Wallaroo admins are provided observability tools to track Wallaroo infrastructure resources across the cluster and by pipelines via the Wallaroo Admin Dashboard and the Wallaroo MLOps API. This empowers administrations with the information they need to observe usage trends, optimize cloud spend, and ensure no operational interruptions.
  • Metrics Admin
  • Wallaroo MLOps API Essentials Guide: Wallaroo Dashboard Metrics Retrieval