TAGGER tag

TAGGER is a full-stack AI data-solutions platform that combines blockchain with data collection, labeling, management, authentication, and trading. Its documentation describes a cross-border, permissionless protocol intended to address data silos, inconsistent data authentication, and the shortage of professional annotators.

The platform uses Web3 crowdsourcing and a proof-of-human-work model to let participants collect, clean, annotate, review, and otherwise process data. An AI Copilot is intended to assist workers across image, voice, text, and video tasks while standardized checks evaluate submitted work.

TAGGER also proposes a decentralized AI-data marketplace. Owners can sell datasets or authorize their use, while blockchain-recorded data-rights certificates are intended to provide traceability of ownership, origin, and usage history. The project positions this infrastructure as a bridge between decentralized networks and AI data production.

TAGGER targets the fragmented data economy underlying AI development. Its stated challenges include data silos, difficult data acquisition, inconsistent or chaotic authentication, low-quality labels, privacy and ethical risks, and the continuing maintenance burden for datasets. These issues can make data expensive to source, difficult to verify, and hard to share across borders.

The project argues that centralized data ownership and opaque intermediaries limit participation and create uncertainty about who may use a dataset. Its proposed response is cryptographic and smart-contract-based authentication, decentralized collection and sharing, AI-assisted annotation, and a permissionless marketplace that records transfers or usage authorizations on-chain. These are project goals and design claims; the documentation does not establish that every proposed service is already operational.

Data owners upload data and declare ownership through TAGGER's data-authentication workflow. The documentation says data is encrypted using deep-learning-related techniques and stored or transmitted through a decentralized network; unique data-rights certificates represent control, allowing the owner to manage viewing, trading, or authorization. Blockchain records are intended to make a dataset's origin, ownership, and usage history traceable.

Participants perform collection, cleaning, annotation, RLHF, and review tasks. The AI Copilot supports workers and a standardized verification process evaluates results. The published reward formulas multiply a base task reward by a halving coefficient, account-level coefficient, and task-specific quality coefficient: mask accuracy for AI-Copilot labeling, daily pass proportion for manual labeling, and daily review accuracy for data review. Account coefficient is documented as 1 + 0.01 × account level.

$TAG is described as the native and governance token. It can be used for posting tasks, staking, buying or using datasets, software subscriptions, AI-model customization, and marketplace fees. The stated supply is 405,380,800,000 TAG: 300 billion (74.00449158%) for proof-of-human-work, 85.3808 billion (21.06187565%) for Tag-to-Pump, and 20 billion (4.93363277%) for liquidity. The documentation describes 5% task-listing fees for collection, cleaning, and annotation orders, and 1% listing/handling fees for dataset sales or rentals.