---
title: "Grabette: An Open System for Recording Robot Manipulation Data"
source: "Hugging Face：Blog（RSS）"
source_tier: "t1"
url: "https://huggingface.co/blog/grabette"
permalink: "https://feed.astralor.com/items/itm_f07ea0986edc9c353236b2f4"
published: "2026-07-21T00:00:00.000Z"
score: 85
channel: "ai"
tags: ["Grabette","Hugging Face","Open-source System","Robotic Manipulation"]
---

# Grabette: An Open System for Recording Robot Manipulation Data

## AI commentary

> Grabette fills a critical gap in the toolchain for robot learning data collection. For researchers in embodied AI, imitation learning, or reinforcement learning, this open system can significantly reduce the non-core burden of data recording. I particularly appreciate its direct integration with Hugging Face Datasets, which could foster standardized dataset publishing and community collaboration.

## At a glance

Hugging Face team releases Grabette, an open system designed to streamline the recording and collection of robot manipulation data. It supports multiple sensor inputs, provides standardized data formats, and integrates visualization and replay tools to help researchers efficiently build robot learning datasets. Grabette's modular design makes it easy to extend and compatible with mainstream robot platforms, potentially advancing open and reproducible embodied AI research.

## Key points

- Grabette provides a unified interface for recording robot manipulation data with multi-sensor support.
- Built-in visualization and replay features enable data quality checks and annotation.
- Modular architecture adapts to different robot hardware and software environments.
- Open-source release fosters openness and collaboration in robot learning.

## Full text

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_f07ea0986edc9c353236b2f4