Bringing Guardrails in Production for AI Apps using Portkey and Postman

Presented by Vrushank Vyas, Dev Rel at Portkey.

LLM outputs are still unreliable. Some of it is fixed with having strong AI guardrails that scan your inputs & outputs and fix LLM behavior on the fly. By using Postman along with Portkey's open source AI Gateway, we now bring these Guardrails right on top of the Gateway to help AI teams fully orchestrate their requests EXACTLY the way they want. This talk will demonstrate how the "Guardrails on the Gateway" pattern is crucial for all AI teams going to production.