Tools/Explorers

Blockchain Data Indexers Compared: The Graph vs Goldsky vs SubQuery

Compare blockchain indexing services by query speed, chain coverage, pricing, and decentralization model. The Graph, Goldsky, SubQuery, Envio, Ponder, and Shadow.

Spark Team

Overview

A blockchain indexer transforms raw on-chain data into structured, queryable databases. Without one, reading historical contract events means scanning every block sequentially: a process that can take hours or days for a mature chain like Ethereum. Indexers solve this by pre-processing blockchain data and exposing it through APIs (typically GraphQL or SQL), enabling sub-second queries over terabytes of on-chain history.

The indexer market has fragmented rapidly. The Graph pioneered the subgraph pattern and operates a decentralized network of indexers. Goldsky built a managed platform around subgraph compatibility and real-time streaming. SubQuery expanded from Polkadot into 280+ networks. Envio focused on raw speed with its HyperSync engine. Ponder offers an open-source, self-hosted TypeScript framework. Shadow takes a different approach entirely, using off-chain contract execution to generate custom event data without gas costs.

The table below provides a high-level comparison. Each indexer is examined in depth throughout this guide.

IndexerTypeChain CoverageQuery InterfacePricing ModelSelf-Hostable
The GraphDecentralized network40+ chainsGraphQLPer-query (GRT or USD)Yes (graph-node)
GoldskyManaged platform150+ chainsGraphQL, SQL sinksUsage-basedNo
SubQueryDecentralized network280+ networksGraphQLSQT token / managed tiersYes
EnvioManaged + self-hosted83+ EVM (HyperSync)GraphQL, SQLTiered (free dev tier)Yes
PonderOpen-source frameworkAny EVM (via RPC)GraphQL, SQLFree (self-hosted)Yes (only option)
ShadowManaged platformEthereum, Arbitrum, Optimism, BaseSQL (Postgres), JSON-RPCCustomNo

Indexing Speed and Performance

Historical backfill speed determines how quickly a new deployment can serve queries. An independent benchmark by Sentio (April 2025) tested indexing the Uniswap V2 Factory contract on Ethereum mainnet:

  • Envio (HyperSync): ~8 seconds
  • SQD (Subsquid): ~2 minutes
  • The Graph (Subgraph Studio): ~19 minutes
  • Ponder: ~21 minutes

Goldsky was not included in this benchmark because it uses subgraph-class indexing under the hood. Shadow was excluded because it operates on a fundamentally different model (off-chain contract execution rather than event log indexing).

Envio's speed advantage comes from HyperSync, a purpose-built data retrieval layer that bypasses standard RPC endpoints and reads directly from optimized chain snapshots. The Graph's Substreams product offers significantly faster processing than standard subgraphs, but requires writing Rust modules rather than the familiar AssemblyScript mapping handlers.

For real-time query latency after initial sync, most hosted services deliver sub-100ms responses for typical GraphQL queries. The Graph's decentralized network adds routing overhead compared to centralized alternatives, but query latency is generally acceptable for most dApp frontends.

Chain Coverage

Chain coverage varies dramatically across indexers. SubQuery leads in raw network count, while Envio and Goldsky focus on EVM-compatible chains with optimized performance. The Graph sits in between, covering 40+ chains on its decentralized network including several non-EVM ecosystems.

EcosystemThe GraphGoldskySubQueryEnvioPonderShadow
Ethereum + L2sYesYesYesYesYesYes
SolanaYesYesYesNoNoNo
Polkadot / SubstrateNoNoYes (59 parachains)NoNoNo
CosmosYesNoYes (49 chains)NoNoNo
NEARYesNoYesNoNoNo
Bitcoin L2s (EVM)NoVariesYes (BOB, Bitlayer, BEVM)Yes (if EVM)Yes (if EVM)No
StellarYesNoYesNoNoNo

For teams building on Bitcoin L2s, the choice depends on the L2's architecture. EVM-compatible Bitcoin L2s like Botanix, BOB, and Bitlayer can be indexed by SubQuery, Envio, or Ponder using standard EVM indexing. Non-EVM Bitcoin protocols require specialized indexers. For native Bitcoin UTXO data, tools like Maestro's Symphony or Esplora are more appropriate than general-purpose subgraph-style indexers.

Bitcoin Layer 2 solutions like Spark use off-chain state models that differ from both UTXO scanning and EVM event log parsing, so developers should check each L2's documentation for recommended data access patterns.

Pricing Models

Indexer pricing falls into four categories: per-query fees, usage-based metering, flat subscription tiers, and self-hosted (where you pay for your own infrastructure).

The Graph's Subgraph Studio offers a free tier of 100,000 queries per month. Beyond that, additional queries cost $2 per 100,000, payable by credit card or in GRT tokens. On the decentralized network, query pricing is set by individual indexers competing in an open market. Enterprise plans with custom pricing are available for high-volume consumers.

Goldsky uses usage-based pricing with a Starter tier that includes a one-time $100 credit and 1 million monthly Edge RPC requests. Its Scale tier is pay-as-you-go with unlimited subgraphs, pipelines, and team members. Envio offers tiered pricing from a free development plan through production options via Envio Cloud. Ponder is free and MIT-licensed: the only cost is running your own Postgres database and RPC endpoints.

SubQuery's decentralized network uses SQT tokens for payments between consumers and node operators. Pricing is market-driven, with the network designing incentive structures to balance consumer costs against operator returns. SubQuery also offers an enterprise managed service for teams that prefer not to interact with the token economy directly.

Feature Comparison: Streaming, Resolvers, and Reindexing

Beyond basic event indexing, production data pipelines require capabilities like real-time streaming, custom transformation logic, and the ability to reindex historical data when schemas change.

FeatureThe GraphGoldskySubQueryEnvioPonderShadow
Real-time streamingSubstreams (Rust)Mirror (CDC)NoYes (live mode)Yes (hot reload)Yes (live sync)
Custom resolvers / transformsAssemblyScript / RustSQL transformsTypeScriptTypeScript / ReScriptTypeScriptSolidity (shadow contracts)
Historical reindexingFull redeployAutomaticFull redeployAutomaticAutomatic (dev server)Full redeploy
Multi-chain aggregationSeparate subgraphsYes (pipelines)Yes (single project)Yes (multi-chain config)Yes (config-driven)No
Data sinksGraphQL onlyPostgres, ClickHouse, Kafka, S3, ElasticsearchGraphQL, PostgresPostgres, ClickHouse (experimental)PostgresPostgres, Dune
Wildcard event filteringNoNoNoYesYesN/A
Reorg handlingAutomaticAutomaticAutomaticAutomaticAutomaticAutomatic

Shadow's approach is unique: rather than indexing existing event logs, it lets developers write custom Solidity functions that execute against live contract state without paying gas. The output is synthetic event data that can expose internal variables, computed metrics, or aggregations that the original contract never emitted. This makes Shadow complementary to traditional indexers rather than a direct substitute.

Goldsky's Mirror product deserves special attention for teams building analytics pipelines. It functions as a change-data-capture (CDC) stream that pushes indexed data to external databases like ClickHouse, Kafka, or S3 in near real-time, enabling teams to run complex analytical queries outside the constraints of a GraphQL API.

Decentralization Tradeoffs

The Graph and SubQuery operate decentralized networks where independent node operators run indexing infrastructure and earn token rewards for serving queries. This design eliminates single points of failure and vendor lock-in, but introduces tradeoffs in latency, pricing predictability, and developer experience.

On The Graph's decentralized network, queries are routed through a gateway that selects among competing indexers based on price, latency, and stake. Query costs fluctuate with the GRT token price and indexer market dynamics. The hosted service, which offered free, centralized indexing, was fully deprecated in 2025, pushing all projects to either the decentralized network, Subgraph Studio, or third-party hosts like Goldsky.

SubQuery's network uses SQT tokens and a delegated proof-of-stake model. Node operators stake SQT to participate, and delegators can stake behind operators to share in query fee revenue. The network is still maturing, with ongoing adjustments to token mechanics and reward structures planned through 2026.

Centralized alternatives like Goldsky, Envio, and Shadow offer more predictable pricing and simpler onboarding. The tradeoff is trust: your data pipeline depends on a single vendor's uptime, data integrity, and business continuity. For applications where verifiability matters (such as proof of reserves dashboards or on-chain governance tools), a decentralized indexer provides stronger guarantees.

How to Choose the Right Indexer

The right indexer depends on your chain targets, performance requirements, team expertise, and tolerance for vendor dependency.

  • Building on EVM chains and need the fastest possible backfill: Envio (HyperSync) offers order-of-magnitude speed improvements over subgraph indexing.
  • Migrating from The Graph's deprecated hosted service: Goldsky provides subgraph-compatible hosting with minimal migration effort, plus streaming and sink features.
  • Indexing non-EVM chains (Polkadot, Cosmos, Stellar): SubQuery is the only option in this comparison with broad non-EVM coverage.
  • Want full control and no vendor dependency: Ponder is free, open-source, and self-hosted. You own the infrastructure and the data.
  • Need custom contract-level data that was never emitted as events: Shadow's off-chain execution model generates synthetic event data from arbitrary Solidity logic.
  • Building for decentralization and censorship resistance: The Graph or SubQuery's decentralized networks eliminate single-vendor risk.

For teams evaluating crypto data APIs more broadly, indexers are one layer in a larger stack that may also include block explorer APIs, oracles, and direct RPC access. See our Bitcoin block explorer API comparison for UTXO-chain-specific options.

Frequently Asked Questions

Do I need a decentralized indexer?

Not necessarily. Decentralized indexers like The Graph and SubQuery provide censorship resistance, redundancy across multiple operators, and verifiable data integrity. These properties matter for public goods infrastructure, governance dashboards, and applications where trust minimization is a core requirement. For internal analytics, prototyping, or applications where you already trust a single cloud provider, a centralized managed service is simpler and often faster.

Which blockchain indexer supports Bitcoin data?

None of the general-purpose indexers in this comparison natively index Bitcoin's UTXO-based chain. SubQuery supports several EVM-compatible Bitcoin L2s including BOB, Bitlayer, and BEVM. Envio and Ponder can index any EVM-compatible Bitcoin L2 through standard RPC configuration. For native Bitcoin data (transactions, Ordinals, Runes, BRC-20), purpose-built indexers like Maestro Symphony, Esplora, or Electrum server are more appropriate. See our API economy and payments infrastructure research for how indexing fits into the broader data access landscape.

What is the difference between a subgraph and a custom indexer?

A subgraph is a declarative indexing specification popularized by The Graph. You define a schema, specify which contract events to listen for, and write mapping handlers (in AssemblyScript or TypeScript) that transform events into entities. A custom indexer is any purpose-built data pipeline: it might use an RPC node, a direct database connection, or a specialized engine like Envio's HyperSync. Subgraphs offer portability (they can run on The Graph, Goldsky, or self-hosted graph-node), while custom indexers offer more flexibility at the cost of vendor-specific implementation.

How much does blockchain indexing cost at scale?

Costs vary widely. The Graph charges $2 per 100,000 queries beyond its free tier of 100,000 monthly queries. At 10 million queries per month, that is roughly $200. Goldsky's usage-based pricing scales with compute hours and data volume rather than query count. Self-hosted options like Ponder shift the cost to your own infrastructure: a production Postgres instance, RPC endpoint fees, and server resources. For high-volume applications, the cost of RPC calls to source chain data often exceeds the indexer's own fees.

Can I migrate between indexers?

Migration difficulty depends on how closely your current indexer follows the subgraph standard. Subgraphs deployed on The Graph can often run on Goldsky with minimal changes. Moving from a subgraph to Envio or Ponder requires rewriting mapping handlers in TypeScript, though the schema and event selection logic are conceptually similar. Moving to or from Shadow requires a fundamentally different approach, since Shadow uses Solidity-based off-chain execution rather than event log processing.

What is HyperSync and why is it faster than RPC-based indexing?

HyperSync is Envio's proprietary data retrieval layer. Traditional indexers fetch blockchain data by making sequential JSON-RPC calls to a node, which is bottlenecked by network latency and node I/O. HyperSync bypasses this by reading directly from pre-processed, columnar chain snapshots optimized for bulk data retrieval. In benchmarks, this approach indexed the Uniswap V2 Factory contract in approximately 8 seconds compared to roughly 19 minutes via standard subgraph indexing on The Graph.

Which indexer is best for real-time data streaming?

Goldsky's Mirror product is the most mature option for streaming indexed data to external systems. It supports change-data-capture to Postgres, ClickHouse, Kafka, S3, Elasticsearch, and other sinks. Envio also supports live-mode indexing with low latency. The Graph's Substreams product offers real-time processing but requires Rust modules and is more complex to set up. For simpler use cases, Ponder's development server provides hot-reload indexing that stays current with the chain tip.

This tool is for informational purposes only and does not constitute financial or technical advice. Data is approximate and based on publicly available information as of late 2026. Pricing, chain coverage, and features change frequently. Always verify current details on each provider's official documentation before making infrastructure decisions.

Build with Spark

Integrate bitcoin, Lightning, and stablecoins into your app with a few lines of code.

Read the docs →