All products
Intelligence

OhhO Autonomy

Map it, navigate it, command it in plain language.

Teleop is for when a human drives; Autonomy is for when the robot drives itself. It maps a space with SLAM, navigates it with Nav2, and runs a mission planner that sequences navigation and manipulation — all orchestrated by a natural-language agent that turns an instruction into a plan.

At a glance
  • 2-D & 3-D SLAM mapping
  • Nav2 path planning + obstacle avoidance
  • Mission planner (navigate → act sequences)
  • Natural-language agent (LLM-backed, your choice of model)
  • Safe control-mode mux: nav / AI / teleop
LIVE
This isn't a mockup — open mission control
Overview

What you get

Perception tells a robot what's around it; Autonomy decides what to do about it. OhhO Autonomy is the behavior layer that turns a powered-on robot into one that moves through the world and completes tasks on its own.

It builds and localizes against a map with SLAM (2-D and 3-D), plans collision-free paths with Nav2 over a fused costmap, and fuses wheel odometry and IMU through an EKF for reliable pose. A mission planner sequences higher-level jobs — navigate to a named location, then run a manipulation policy — and a control-mode mux arbitrates cleanly between navigation, AI policies and human teleop so they never fight over the wheels.

On top sits a natural-language agent, backed by the LLM of your choice, that turns 'take the red cup from the kitchen to the bench' into a structured mission: it knows your named locations, can describe what it sees, and asks for clarification when an instruction is ambiguous — then hands the mission to the planner to execute.

Capabilities

Built to do the hard parts for you

SLAM mapping

Build and localize against 2-D and 3-D maps; switch between mapping and localization modes against a saved map.

Nav2 navigation

Collision-free path planning and obstacle avoidance over a fused costmap, tuned to the robot's real velocity and acceleration limits.

Robust localization

An EKF fuses wheel odometry and IMU so pose stays trustworthy even when mecanum wheels slip.

Mission planner

Sequence navigation and manipulation into a mission — 'go to the kitchen, then pick up the cup' — as a tracked state machine.

Natural-language agent

An LLM-backed agent maps plain-language instructions to missions, with named locations, scene description and clarification when it's unsure.

Safe arbitration

A control-mode mux switches between navigation, AI policies and teleop so commands never conflict — and a human can always take over.

Workflow

How it works

1
Map the space

Drive once while SLAM builds a map, then save it for localization.

2
Name the places

Tag locations — kitchen, dock, bench — the planner and agent can refer to.

3
Give an instruction

Type or speak a task; the agent turns it into a structured mission.

4
Watch it execute

Nav2 and the policies run the mission; the mux keeps teleop override one tap away.

Under the hood

Specifications

Mapping
2-D & 3-D SLAM (mapping / localization)
Navigation
Nav2 planner + costmap obstacle avoidance
Localization
EKF fusing wheel odometry + IMU
Missions
Navigate → manipulate state machine
Agent
Natural language → mission (LLM-backed)
Arbitration
Control-mode mux: nav / AI / teleop
Pricing

Which plan do I need for OhhO Autonomy?

We recommend the Builder plan. Builder gives you mapping, navigation and the mission planner on real hardware — the core of autonomy. Add Fleet when you want the natural-language agent and reusable named missions; Forge is for custom behaviors and running the agent on-prem.

Spark
Freeforever
SLAM + navigation in simulation
Start free
Fleet
$199/ mo
+ natural-language agent & named missions
Choose Fleet
Forge
Customcontact us
Custom behaviors + on-prem agent
Talk to us
FAQ

Common questions

How is this different from OhhO Pilot?

Pilot is for a human operating the robot; Autonomy is for the robot operating itself. They share the same safe control-mode mux, so you can hand control back and forth instantly.

Does the language agent need the cloud?

By default it calls a cloud LLM, but the agent layer is optional and model-agnostic — navigation and the mission planner run fully on-robot, and Forge can run the agent on-prem with your own model.

The platform

Works better together

Ready to build with OhhO Autonomy?

Start free and simulate first — no hardware required. Upgrade when you're ready to deploy.