The Grass infrastructure consists of two key components:
Grass nodes allow users to contribute their unused internet bandwidth to the network, earning rewards and participating in the growth of AI.
The Sovereign Data Rollup consists of the network of nodes, routers, validator, ZK processor, and Data ledger which facilitates the data sourcing and transformation process — turning unstructured web data into structured datasets.
Grass is building the first-ever Sovereign Data Rollup. It streamlines data sourcing and transformation through a globally distributed network of Grass nodes, enabling AI universal access to structured web data.
Validators receive, verify, and batch the Router’s web transactions. They then generate ZK proofs to checkpoint session data on-chain. The on-chain proof can be referenced in datasets to verify data provenance and track its lineage throughout its lifecycle. The validator set will transition from an initial centralized framework with a single validator to a decentralized committee of validators.
Grass Routers connect Grass Nodes to the Validator. The routers keep the network of nodes accountable, and relay bandwidth. They are incentivized to operate and earn rewards as a proportion to the total validated bandwidth served through its relay.
Grass Nodes leverage the user’s underutilized bandwidth and relay traffic so the network can scrape public web data (not your personal data). Running a node is free and easy. Those who run nodes (node operators) are compensated for the data relayed through them.
Zk Processor
The ZK Processor batches validity proofs of session data for all web requests and submits the proofs on a layer 1 blockchain. This action creates a permanent record of every act of scraping ever performed on the network. Furthermore, this lays the groundwork for total visibility into the provenance of AI training data.
Grass Data Ledger
The Grass Data Ledger is the link between the data being scraped and the L1 Settlement Layer. The ledger is an immutable data structure that hosts the full datasets and links the data to its corresponding on-chain proof. Effectively, it is the data repository that ensures data provenance.
Edge Embedding Models
Edge embedding is the process that converts unstructured web data into structured models. This encompasses all the necessary pre-processing steps, ensuring that the raw data collected is cleaned, normalized, and structured in a format that aligns with the requirements of AI models.
