Getting started

Accounts and roles

Sign up at ay-robots.com with an email address, Google, or GitHub. You need to verify your email before the platform lets you in. During onboarding you choose a role, and that choice decides which dashboard and features you see afterwards.

Complete onboarding

Clients add company information and billing details, then move on to creating an environment and pairing a robot, which the rest of this page covers.

Operators fill in country, timezone, an hourly rate, and a short bio, then run the built-in speed test to confirm the connection is good enough for low-latency control. Payouts go through Stripe Connect, but you can finish that part later; it only becomes necessary once you have earnings to withdraw. The operator guide covers certifications and payouts in detail.

Create your first environment

An environment is a workspace that groups robots, cameras, sessions, and the datasets they produce. The practical rule: one environment per physical rig or location. If you have one SO-100 on a desk in the office, that desk is your environment. Create it from the dashboard, give it a name you will recognize in a list, and everything else you set up will live inside it.

Keeping environments aligned with physical reality pays off later. Camera names, robot roles, and session history all read much better when an environment corresponds to one actual table you could point at.

Pair your robot

Pairing links a physical arm to your environment using a short one-time code. The code is generated in the browser and redeemed on the machine that is physically connected to the arm, either the desktop client or the robotapp on a Raspberry Pi.

1. 1
   Start the pairing wizard

   In your environment, choose to add a robot. The wizard generates a short pairing code and waits for it to be redeemed.

2. 2
   Redeem the code at the arm

   On the computer connected to the arm, open the desktop client (or the robotapp on the Pi) and enter the code. The code is single-use and links exactly this hardware to exactly this environment.

3. 3
   Wait for the heartbeat

   Once the code is redeemed, the robot starts sending heartbeats to the platform. The wizard switches to connected, and the robot appears as online in the environment.

4. 4
   Assign the cameras

   Detected cameras are listed next to the robot. Assign each one and give it a name that describes its viewpoint, such as wrist or scene. These names travel into your datasets, so pick them once and keep them.

5. 5
   Test a small motion

   Open the Control tab and send a few small joint impulses. If the arm responds and the camera feeds update, pairing is done and the robot is ready for sessions.

Run your first session

With a robot online, book a session from the environment, or leave the robot available for a certified operator to pick up. The session is created as PENDING and turns ACTIVE when the operator connects. While it runs, you can watch the camera feeds live, follow the activity events, and talk to the operator in the session chat.

When the session ends, it moves to COMPLETED. You can rate the operator, and the recording shows up as episodes in the environment's dataset browser. That first dataset is worth opening even if it is small: seeing the video next to the joint charts makes it obvious what the platform actually captured, and what a training pipeline will later consume.

From here, the client guide covers day-to-day work with robots and datasets, the sessions page explains the lifecycle and events in detail, and the teleoperation page documents the cockpit controls.

Frequently asked questions