Muted demo loop from the projectCredit: Dobb·E project page (NYU)

Teach household tasks with Dobb-E and a reacher-grabber

Wheeled, Hard

Dobb-E collects home demos with a reacher-grabber tool and an iPhone, then trains policies that transfer to a Stretch robot with little new data.

View build on GitHub ↗Share on X

From NYU's 'On Bringing Robots Home' paper: open-source hardware, the Homes of New York dataset, and pretrained models.

Hardware & cost: A reacher-grabber, an iPhone, printed mounts, and a Stretch robot to deploy.

Difficulty: Hard.

Agent install prompt (paste into Claude Code, Cursor or Codex): "Read https://github.com/notmahi/dobb-e and docs.dobb-e.com, and list what I need to print and install to collect my first demonstrations."

Source: notmahi/dobb-e on GitHub — https://github.com/notmahi/dobb-e (facts checked against the project README, Oct 2026; prices vary by region).

Quick answers

What do you need to teach household tasks with Dobb-E and a reacher-grabber?: A reacher-grabber, an iPhone, printed mounts, and a Stretch robot to deploy.
How hard is it to teach household tasks with Dobb-E and a reacher-grabber?: Hard.