Skip to content

Getting Started

PhysiCar AI is a creative Physical AI education platform.
Through hands-on examples, you build your own robot and self-driving with your own hands and grow into a Physical AI Creator.

Anyone can start — free, easy, and fast

Every hands-on activity runs in your web browser, with nothing to install — we recommend the Chrome browser.

You can start for free with the $5 welcome credit that comes with your PhysiCar AI account (launch event).

About PhysiCar AI

Platform structure

PhysiCar AI is made up of three parts.

  • Site (physicar.ai)

    The platform's official homepage.
    Manage your account and learn and grow on your own through tutorials, learning docs, and more.

  •   Simulator (SIM)

    A space to practice with a virtual robot in the simulator.
    Practice right away with a virtual robot, with no real robot.

  •   Physical robot (REAL)

    The physical robot (Real).
    After practicing with a virtual robot in the simulator, optimize on the real robot.

Sim2Real

PhysiCar AI provides a digital twin practice environment. The simulator has the same web console, ROS 2 topics, and Web API as the real robot, so the code and models you practice on a virtual robot — even without a real robot — run as-is on the physical one.

Build it virtually → verify and optimize it on the real robot: this whole flow is Sim2Real.

Concept → Digital twin · Start now → Start with a virtual robot

Physical robot (REAL)

PhysiCar is physical + car — a 1/16-scale real robot that combines a car-shaped body with AI.

Part Spec
Computer Raspberry Pi 5
Camera 480×360, pan/tilt ±30°, night vision
LiDAR 360°, approx. 0.1–16 m
IMU 6-axis (accelerometer + gyroscope)
Drive Ackermann steering ±20°, up to 3.0 m/s
Other Speaker, rear touchscreen

Structure & sensors in detail → Robot anatomy & sensors

Your learning path

Start easy, follow along, and let the "why" wait for the Learn pages.

  1. Sim2Real — run a virtual robot in your browser, then a real one, then move your work between them.
  2. Basics — the workspace, the robot's controls and sensors, the AI tutor, and your account & credits.
  3. Racing — one track, every AI method: rule-based → supervised learning → reinforcement learning.
  4. Agent — turn an LLM into a robot agent with your own instructions and tools, then vibe-code a project.
  5. Advanced — object detection, navigation (SLAM · Nav2), your own robot app, and the APIs.
  6. World Building — design a world, add an evaluation, publish it, and print the real track.

Teachers & organizations

  • Classroom operation — students join with an invite code; their usage is billed to the teacher's credits within limits you set (classroom · billing).
  • Classes & competitions — publish worlds to your classroom, auto-score with evaluations, and run offline events with printed tracks and the Racing Lap Timer (guide).

Get ready

Create your account

You'll need an account for a smooth experience. AI chat and simulator usage time are deducted from your credits.

  • Sign up for PhysiCar AI — Sign up (email-based). Your PhysiCar AI account is all you need — no GitHub account required.

Credits & pricing in detail → Credits & pricing

Buy a physical robot (optional)

If you want to practice with a real robot, buy the robot, official track, fences, and accessories from the Store.

Once it arrives → Start with a physical robot

AI