The robots are learning. You're at the controls.

Physical AI is trained on human demonstrations. StreamGenie staffs the operators who make them, teleoperating robot arms and running instrumented capture sessions for robotics labs. Remote and in-lab work, by application.

  • No fees, ever
  • Paid trial task
  • Rates shown before you accept
Recording
Leader arm You move it by hand, or remotely
Follower arm The robot mirrors every motion
Each attempt becomes a demonstration episode recorded in the robot's own body: the data robotics labs train on.

Pick the work that fits you

Two operator tracks by application, and one you can start tonight.

Teleop operator

Drive robot arms through real tasks (pick, place, fold, sort) so the robot learns from demonstrations recorded in its own body. You work a leader arm; the robot's follower arm mirrors you.

Where
Remote or in-lab
Work
Scheduled sessions
Pay
Set per project
Apply as a teleop operator

Capture operator

Run instrumented capture sessions to a buyer's spec: performing tasks in the POD, our capture station (an egocentric camera plus an external view), and keeping every take clean and on-protocol.

Where
In-lab, or your space to spec
Work
Session blocks
Pay
Set per project
Apply as a capture operator

Start today with a campaign

Not ready to apply? Take on a paid marketplace campaign: record real-world tasks or cut clips to a published brief, on your own time. It's the fastest way to get paid for real-world data, and a track record we look at.

Where
Wherever you are
Work
Whenever you like
Pay
Shown on every brief
Browse campaigns

Careful, repeatable, and it matters.

Robotics labs can't train a robot to fold a towel from text on the internet. They need thousands of clean demonstrations of the real task, and the most useful ones are recorded in the robot's own body, by a person driving it. That person is an operator.

It isn't gig-app busywork: sessions are scheduled, specs are precise, and quality is checked before anything ships.

  1. 01
    Brief Every session starts with a task spec: what to demonstrate, what good looks like, and what gets a take rejected.
  2. 02
    Demonstrate You perform the task through the leader arm, or in the POD: smooth, deliberate, at human speed. Quality beats volume.
  3. 03
    Episode Each attempt is saved as one episode. Fumbles are fine; you flag them and go again.
  4. 04
    QA Episodes are checked against the spec before anything ships. Consistent operators get more sessions.

You don't need a robotics degree

The best operators share a temperament more than a résumé.

Steady and patient Long sessions of the same motion, done carefully every time.
Follows a spec to the letter The buyer's protocol is the job. Precision beats improvisation.
Reliable You show up for the sessions you book. Teams are built around that.
Good with your hands Gaming, trades, hobby robotics, sim rigs, instruments: it all transfers.

From application to your first session

  1. 01 Apply Three short steps, about three minutes. We read every application.
  2. 02 Short screen A quick call or async check-in about your experience, setup and availability.
  3. 03 Paid trial task A small, real task so you can see the work and we can see your quality. It's paid.
  4. 04 Matched to projects We match you to sessions as buyers commission them. Rates are shown before you accept.

Pay, plainly.

Operator rates are set per project (by task, hardware and location), and you see the rate before you accept any session. Payouts go to PayPal, the same as every campaign on the marketplace. No fees, no paid training, no equipment you have to buy from us.

See what the marketplace pays

The AI data economy, talked out loud.

The Lab, our vidcast, is where we unpack what we're building: how AI training actually works, why labs pay for everyday footage, and where creators fit into the supply chain. In 05 we sit down with the guy who buys AI training data.

The Lab 05: The Guy Who Buys AI Training Data

Apply to the operator workforce

Three short steps, about three minutes. Tell us which work fits, what you bring, and how to reach you. We read every application.

  • Teleop and capture tracks: pick one or both
  • Paid trial task before regular sessions
  • Matched to projects as labs commission them
  • Rates shown before you accept anything

Already have a StreamGenie account? Sign in and we'll prefill this and track your application on your dashboard.

  1. 1 Tracks & time
  2. 2 Your background
  3. 3 About you

Which work fits you?

Tracks · pick one or both
Hours you could give each week
1 / 3

Straight answers