Introducing API v5, the agent-first memory API - container tags are now namespaces, and they live in the url - output from the API is now optimized for agent context - one prompt to migrate - Markdown negotiation in API results

Check it out:

Agent-first memory is a big step. But I think the harder layer starts after retrieval: Where did this memory come from, why should the agent trust it, and what changed it? Memory gives continuity. Provenance makes that continuity trustworthy. 🤷🏻‍♂️

Memory APIs get useful once the output fits the context window.

Sick

Love the agent first focus namespaces in the URL and markdown negotiation is exactly the kind of clean upgrade agents need!

one prompt to migrate is doing a lot of heavy lifting

is there a "guide", or something like that, for it to integrate well with Orca? @orca_build

@orca_build We should work on one!

A cleaner API design that makes persistent memory easier for agents

the agents now have opinions about Markdown dialects