The Graph grt

The Graph is a decentralized blockchain-data infrastructure protocol. It extracts, processes, and delivers structured blockchain data so applications do not have to build their own indexing systems. Its original product is the Subgraph: an open, application-specific data API that describes which contracts and events to index, transforms them into entities, and exposes the result through GraphQL.

The protocol has expanded beyond Subgraphs into a suite of data services. Substreams provides parallelized, high-throughput streaming for low-latency indexing and large backfills, while Amp is positioned as an enterprise-oriented blockchain database with SQL access, audit trails, and verifiability. The services are intended for dapps, analysts, AI agents, and institutional teams.

GRT is The Graph Network's native work-utility token. Network participants use it to coordinate data supply and demand: Indexers operate infrastructure and serve data, Delegators delegate stake to Indexers, Curators signal high-quality Subgraphs, and Developers build and query data products. The Graph documentation says the network spans 60+ networks and, as of early 2026, had served more than 1.27 trillion queries to over 75,000 projects.

Blockchains are optimized for writing and verifying transactions rather than reading them. They expose low-level RPC methods and store information in append-only blocks, contract storage slots, and event logs instead of relational tables. Answering a request such as all transfers for a user can require scanning blocks and reconstructing application-level meaning.

Historical queries may require expensive archive nodes, and application data is usually fragmented across contracts and events. Without a shared indexing layer, every dapp must operate and maintain its own data pipeline, database, API, and reliability infrastructure. The Graph addresses this repeated engineering burden with reusable, queryable data products served by a distributed Indexer network.

A developer defines a Subgraph manifest specifying contracts, events, data sources, and mappings. Graph Node follows the manifest, reads chain data, transforms it into entities, stores those entities, and serves them through GraphQL. A published Subgraph can be discovered in Graph Explorer, queried in a playground, and used as an API by dapp frontends. Substreams uses a different parallel streaming model when throughput and latency matter more than an opinionated entity API; Amp targets SQL, auditability, and enterprise workflows.

Indexers run independent hardware and software, stake GRT, select deployments to index, and serve queries. Curators deposit GRT into a Subgraph's bonding curve to create curation shares and signal that it deserves indexing; query-fee participation is tied to those shares. Delegators delegate GRT to Indexers and receive a share of rewards and query-fee revenue according to the Indexer's parameters. Developers or end users pay query fees in GRT for paid network data, while a portion of query fees is burned.

Indexing rewards come from protocol issuance and are distributed according to the deployment's curation signal and the GRT allocated to it. Indexers submit Proofs of Indexing (POIs), cryptographic proofs that a deployment was indexed correctly through a given block. Under the live Graph Horizon design, allocations can remain open indefinitely and rewards are collected through periodic POIs; a POI older than the current 28-day max staleness can allow any participant to force-close an allocation. Dispute mechanisms, slashing, and Fishermen provide incentives against incorrect indexing or query service.