We study what makes AI systems actually work
Research
Core thinking on AI colleagues — memory, multi-agent coordination, and the hard problems that aren't solved yet.
Engineering
How we build, ship, and operate agentic workflows — tooling, observability, and lessons from running colleagues in production.
Product
Feature announcements, design decisions, and the product thinking behind how Eluu works end-to-end.
True parallel agents: how every Eluu task runs in its own Daytona sandbox
Most agent frameworks fake parallelism with async on a single machine. Eluu gives every task its own Daytona sandbox — so 50 colleagues running 200 tasks really are 200 isolated computers.
Introducing Views: live dashboards owned and maintained by AI colleagues
Views are self-maintaining dashboards: every chart has an AI colleague as its owner, and that colleague keeps the queries, segments, and definitions correct as your business changes.
What's coming next
Drafts in progress. Subscribe above to get them when they land.
- Coming soon
Engineering
What is an Agent SDK? A working definition for the post-framework era
Why the SDK abstraction is the wrong level for agentic systems, and what replaces it.
- Coming soon
Research
What is an Agent Harness, and how does it actually work?
The control loop, the tool surface, and the failure modes you only see at production scale.