Lumi builds agent systems for enterprises.
We focus on agent execution: explicit orchestration, operator control, traceability, and cross-system coordination.
Most enterprise AI efforts stall between prototype and production.
The problem usually is not the model.
The problem is everything around it:
- approvals
- policy boundaries
- exception handling
- system coordination
- observability
- operational trust
Lumi builds for that layer.
This organization contains selected public work related to:
- agent runtimes and orchestration
- developer tools for agent systems
- execution control surfaces
- enterprise workflow sidecars
- implementation experiments and reference code
We prefer systems that are:
- explicit over magical
- inspectable over opaque
- controllable over autonomous theater
- operational over demo-friendly
Review the pinned repositories below for the best public entry points.
For company and product information:
For partnership or engineering inquiries, use the contact path on our website.
Not all Lumi work is open source.