What Is Humanoid Network¶
Humanoid Network (HAN) is a Web3-incentivized, egocentric human video data collection platform for Physical AI.
Any Robot. Any Task. One Network.
The Problem¶
The robotics industry faces a critical bottleneck: training humanoid robots and embodied AI systems requires massive volumes of high-quality human demonstration data, but no efficient, scalable mechanism exists to collect it.
Current approaches are fragmented, expensive, and produce inconsistent quality. Research labs record their own datasets in controlled environments, limiting diversity. Commercial data collection is prohibitively costly at scale.
The Solution¶
Humanoid Network solves this by creating a decentralized, token-incentivized platform where anyone with a smartphone can contribute human motion data. Contributors complete structured "quests" that capture specific movements and tasks, earning $HAN tokens as rewards.
This data is processed through an automated quality assurance pipeline, categorized using the ROBO classification system, and made available to robotics companies, research labs, and AI developers through the HAN SDK.
Key Numbers¶
| Metric | Target |
|---|---|
| ROBO Categories | 7 |
| Data Modalities | Video, depth, pose, audio |
| Chain | Base (Ethereum L2) |
| Token | $HAN |
How It Works¶
- Contributors download the Mecka app and complete quests (structured data capture tasks)
- Quality Assurance validates and scores each submission automatically
- Robotics Companies access curated, structured training data via the HAN SDK
- Token Incentives align all participants through the $HAN token economy
Who Is It For¶
- Contributors: Anyone with a smartphone who wants to earn tokens by capturing human motion data
- Robotics Labs: Teams building humanoid robots, manipulation systems, or embodied AI
- Researchers: Academic groups studying human motion, imitation learning, or sim-to-real transfer
- Enterprises: Companies deploying robots in warehouses, manufacturing, healthcare, and homes