A free, hands-on course. Store a memory, recall it in a new session, and reason over it, then wire the same three calls into a real agent and ship it.
From the concept, to your first working loop, to a production agent.
Start here. Why your agent forgets, what agent memory actually means, and what you can build once it remembers.
Your first success, in minutes. An interactive quickstart: run Hindsight, store a memory, recall it in a new session, and see the whole loop end to end.
The core tasks, then your stack. Store and recall the right things, reason over memory, keep each user separate, steer what's remembered, keep memory fresh, then wire Hindsight into Claude Code, MCP, LangGraph, and CrewAI.
Run it for real. Choose Cloud or self-host, deploy with Docker or Kubernetes, choose your models, monitor and scale, and move data between banks.
Everything else in the course is a variation on these.
Hand it text; it extracts the facts and links the entities for you.
Ask in plain language from any session; matches by meaning, ranked.
Get a grounded answer written from what the bank actually knows.
Free forever. Free Cloud credits to run it live as you learn.