NEXUS-MM Documentation

A world model that learns physics from scratch and tells robots when to stop, slow down, or go.

Quick Start

# 1. Install
pip install nexus-mm-sdk

# 2. Use
from nexus_sdk import NexusRobot
robot = NexusRobot()

# Ask NEXUS for a decision
d = robot.decide("move_forward", "ice", 0.8)
print(d.action, d.risk)  # stop 0.95

# Check backend health
print(robot.health())

Core API

FunctionDescription
robot.decide(action, surface, speed)Robot safety decision
robot.simulate_physics(object, height, surface)Physics simulation
robot.test_friction(surface, weight)Friction testing
robot.test_material(material, height, force)Material testing
robot.plan_route(start, goal, dangers)Path planning
robot.predict_cause(action, target)Cause-effect prediction
robot.can_fit(gap_width, gap_height)Spatial reasoning
robot.predict_stack(objects)Stability prediction
robot.train(predicted, actual)Learn from outcomes

Supported Surfaces

tileicecarpetconcretewet_floormudgravelsandrubbledebrisoil_spillsnowgrassrockywaterash