Courts staff need ordinance summaries streamed from a model that is already deployed in Foundry. What should the staff application do?

Consuming the deployed Foundry model from the courts staff application to stream ordinance summaries is the correct next step when the text model is already deployed. Deployment separates hosting from client integration: staff apps authenticate and call the existing Foundry text deployment, then stream tokens into the summarization UI so clerks see progressive output while reviewing ordinances. That approach avoids rebuilding infrastructure for a capability that already exists and was presumably evaluated in Foundry.

Exporting the deployed Foundry model from the courts staff application to stream summaries uses the wrong operation, because exporting packages or downloads model files for use elsewhere rather than streaming tokens live into the courts staff application while clerks review ordinances. Evaluating the deployed Foundry model from the courts staff application to stream summaries uses the wrong lifecycle step, because evaluation scores quality on test data rather than streaming tokens live into the courts staff application while clerks review ordinances. Fine-tuning the deployed Foundry model from the courts staff application to stream summaries starts retraining to adjust model weights rather than calling the existing deployment, which delays ordinance summaries instead of letting the courts staff application stream from the endpoint directly.

Exam caveat: after a Foundry model is deployed, exam items often test consume-and-stream patterns from line-of-business applications rather than exporting a model that only packages files for use elsewhere, not live streaming. Operational check: from the courts staff app, call the deployed Foundry text endpoint with streaming enabled and confirm ordinance-summary chunks render progressively under proper authentication, authorization, and audit logging.