About this guide

Contents

1. Who it is for
2. How it is made
3. Keeping it current
4. Corrections

Most of what is written about teleoperation data is either a research paper describing one lab's rig or a vendor page describing one company's service. A team planning its first collection needs the part in between: what the rig costs, how many clean episodes an operator produces in an hour, what to check before training and what a usable episode ends up costing. This guide collects the published numbers on each and shows where they came from.

Who it is for

Founders, robot-learning engineers and operations leads at companies that build robots and need demonstrations on their own hardware. We assume the reader already has a robot and a task in mind, and is deciding how to staff and run the collection.

How it is made

Each page is built from primary material: research papers and their project pages, official documentation for the data formats, makers' product pages for prices and official statistics for labour costs. Every figure carries a numbered source with the date we opened it. Where a number is reported by the company that benefits from it, the text says so. The rules are set out on the methodology page.

The cost table on the main guide is a worked model. Its inputs are sourced; its outputs are arithmetic that a reader can redo with their own inputs.

Keeping it current

Prices, pay ranges and format versions change. Every page shows when it was last updated and keeps a short log of material changes. We re-check the priced inputs at least once a quarter.

Corrections

If a figure is wrong or out of date, tell us. We correct verified errors quickly and note the change on the page.

Further reading

• Guide · Throughput and qualityEpisodes per hour and quality checks in teleoperation data collectionTwo numbers decide what a dataset costs: how many episodes an operator records in an hour and how many of them pass. What published collections report on both.
• Guide · StaffingIn-house or hired teleoperators for data collection: the costs comparedOnce the rig works, the operator is most of the cost. The published pay ranges, the true employer cost, the published prices for hired collection and a way to compare them.
• Guide · FormatsRobot data formats compared: LeRobot, RLDS, HDF5 and MCAPThe format decides whether the training job can read what the operators recorded. What each of the four common formats stores, from its own documentation.