Mission-first orchestration
Operators issue one objective. Sannaha converts it into a structured task graph, assigns work by capability, and supervises recovery across multiple robots.
Sannaha
Sannaha is building the intelligence layer that turns a high-level production objective into coordinated sensing, transport, manipulation, and recovery across multiple robotic systems.
Diagnose a production fault, retrieve a replacement part, clear the jam, and verify that the cell has restarted.
What Sannaha does
Most robotics systems are built around a single machine, a single mission, or a brittle integration path. Sannaha is aimed at the coordination layer above that stack: understanding the job, matching it to available capabilities, and keeping work moving when the environment changes.
Operators issue one objective. Sannaha converts it into a structured task graph, assigns work by capability, and supervises recovery across multiple robots.
Motor commands stay in simulator and controller logic, not in the language model. That keeps demos honest and operational behavior traceable.
The first scenario focuses on high-value downtime: diagnose a fault, fetch a replacement part, clear a jam, and verify restart.
First credible demo
The first Sannaha demo is intentionally narrow. It focuses on a high-value operations scenario and proves coordination, not scale theater.
Objective received: Restore Cell B.
Scout robot inspects the cell with simulated sensing.
Coordinator identifies the blocked route and replans.
Transport robot retrieves the spare part from inventory.
Manipulator clears the jam and performs the replacement action.
System verifies restart and closes with Cell B operational.
Why now
A credible multi-robot demo with physics, telemetry, task assignment, and replanning.
A clean control layer for issuing goals, viewing status, and reviewing recovery decisions.
Pilot environments where orchestration can be tested against real operational constraints and partner workflows.
For Hub71 and early partners
The immediate goal is a polished simulation that demonstrates task planning, capability matching, replanning, telemetry, and verified task completion. The longer goal is a control layer for mixed robot fleets in real industrial settings.