Ontos robotics and physical AI products.
Ontos builds tools for machines to understand the physical world. Our work covers robot intelligence, shared machine awareness, physical context and verification of machine-executed work.
Bring physical-world understanding into your robotics stack.
For enterprise teams, we provide a focused briefing on Ontos World, our approach to world models for robotics, and how Sense, P37 Neuro, Verify, and UMP can fit into an existing machine environment.
- World model
Understand spatial context, relationships between objects, changes in the environment, and the physical consequences of an action.
- System integration
Connect machine state and current capability with existing controllers, safety systems, and deployment requirements.
- Technical evaluation
Review your use case, assess technical fit and validation needs, and define a practical next step for your team.
P37 Neuro: Robot intelligence across different bodies
A research and engineering program for robot policies that connect perception, task context, robot-body information and action.
The software platform supports training and evaluation workflows. Learned capabilities remain under research and validation, including simulation and physical robot qualification.
Read the technical overviewSource documentation ↗Sense: Physical context from machine state
Sense turns live machine state into structured physical context and information about a machine’s current capability.
Its public repository documents the implementation, adapters and compatibility information for teams evaluating an integration.
Read the technical overviewSource documentation ↗Verify: Verification of machine-executed work
Independent verification and settlement infrastructure for machine-executed work.
Published as @ontosworld/verify on npm. Review the package documentation for supported workflows and integration requirements.
Read the technical overviewSource documentation ↗UMP: Universal Machine Protocol
A manufacturer-neutral protocol built as a testbed for Ontos world models. It helps robots from different manufacturers share identity, capabilities, current state, and bounded high-level assignments.
UMP is an alpha reference implementation and testbed. Each robot retains its native controller and safety system; UMP does not control motors or replace emergency stops.
Read the technical overviewSource documentation ↗Answers for teams evaluating Ontos
What is Ontos World?
Ontos World is the physical-world reasoning layer within Ontos. It is being developed to help robotics systems represent spaces, objects, relationships, change, and the likely physical consequences of an action.
What is a world reasoning model for robotics?
A world reasoning model is a structured representation of a robot’s environment. It helps a robotics system track what is present, understand how things relate, recognize change, and reason about what may happen next.
How can Ontos work with an existing robotics stack?
An enterprise briefing begins with the robot, available machine state, current controllers, safety boundaries, and deployment requirements. The goal is to identify where Ontos software may fit without replacing the machine’s native control or safety systems.
Who is an enterprise product briefing for?
The briefing is for robotics companies, automation teams, AI researchers, and enterprises evaluating physical AI, world models, machine context, or verification for a real deployment.
Who are these tools for?
Robotics developers, AI researchers and teams integrating machines into physical environments. Each project has its own development status and integration requirements.
How can I request a product briefing?
Visit the Ontos homepage and choose “Request a product briefing”, or email hello@ontos.ws.