Back to Search
I

INDEX

ProjectActive

INDEX is a crowdsourced physical data collection platform and pipeline built by Figure AI. Announced on August 25, 2026 after four months in stealth, INDEX rebrands Figure's data-collection app into an official program for building what Figure describes as the largest and most diverse physical dataset for robot training. The INDEX app pays a global network of contributors ("Creators") to upload videos of everyday real-world tasks such as cooking, cleaning, laundry, logistics, restaurant and factory work, either recorded at home or booked through the app. Figure reports 264,000 app downloads across 108 countries, over 44,000 weekly active users, more than 16 million uploaded videos, 30 minutes of video processed every second, and 15 million USD paid out to Creators to date. Per 1,000 hours collected, INDEX contains 373 unique tasks, 1,146 unique manipulated objects, and 116 unique environments. The data pipeline has five stages: filtering, fraud review, deduplication, rebalancing, and annotation, producing hierarchical text captions for every episode. Figure positions INDEX as the data flywheel behind its Helix models: the official announcement states that the goal is to "collect the physical data Helix needs directly from humans, at scale," and that "Index is our answer to the data problem." (All figures as reported by Figure AI.)

Details

Updated:8/27/2026

No structured details available.

Tags

helixfigure-aidata-collectioncrowdsourcingphysical-datahumanoid

Relationships

Sources

Introducing Index: Building The World's Largest and Most Diverse Physical Dataset
website
Visit
Index | Figure
website
Visit

Appears In

No related landscapes.