Goldsky

Goldsky is a blockchain data infrastructure company founded in late 2021, built on the premise that accessing and processing on-chain data is far harder than it should be. The company provides a suite of tools that let developers index blockchain events, stream them into familiar database destinations, query them through GraphQL APIs, and automate workflows — all without managing the underlying pipeline infrastructure themselves.

Founding and Funding

CEO and co-founder Kevin Li came to the problem from experience in data engineering. Before founding Goldsky, Li worked at Heap Analytics, then at Google and Meta before joining 0x Labs, where he helped build the Matcha decentralized exchange aggregator. That path gave him a concrete view of how difficult real-time on-chain data is to consume correctly. "Real-time data is actually really hard to get right," Li told TechCrunch at the time of the company's seed round, noting that transaction confirmation adds a layer of complexity that most data engineering patterns don't account for.

In September 2022 Goldsky raised a $20 million seed round led by Felicis Ventures and Dragonfly Capital, bringing total funding to $22 million including earlier angel investments. Notable angels included Elad Gil, Plaid co-founders Zach Perret and William Hockey, Uniswap Labs, Zhuoxun Yin of Magic Eden, and Robert Leshner of Compound Finance — a roster that reflects the company's positioning at the intersection of crypto infrastructure and developer tooling.

Core Products

Goldsky's platform has expanded significantly since its founding and now covers five primary areas.

Subgraphs are hosted GraphQL API endpoints that let developers query on-chain data without maintaining indexing infrastructure. Goldsky's subgraph service is fully compatible with The Graph's subgraph specification, meaning teams can migrate existing subgraphs to Goldsky with a single CLI command and zero code changes. Subgraphs handle the indexing automatically, keeping data fresh and queryable at all times.

Mirror Pipelines stream blockchain data directly into a developer's own database. Rather than querying a shared API layer, Mirror lets teams replicate on-chain events in real time to PostgreSQL, BigQuery, ClickHouse, S3, Kafka, or other destinations, giving them full control over how they store and process data. This suits applications that need to run complex analytics or join on-chain data with off-chain records.

Turbo Pipelines build on Mirror's architecture with TypeScript-based transform logic, faster spin-up times, and live data previews. The product allows teams to filter, reshape, and route event data as it flows from the chain into their systems — a capability aimed at production-grade pipelines where raw on-chain data needs to be cleaned or enriched before landing in a database.

Edge RPC is a multi-region RPC endpoint built on open-source foundations (including eRPC) that provides access to EVM networks through a global CDN with automatic failover. The product targets latency-sensitive applications where a single-region RPC provider creates unacceptable risk.

Compose, launched in May 2026, is positioned as an execution layer for onchain and offchain workflows. It allows developers to write durable TypeScript programs that schedule and automate transactions, move money, update smart contracts, or trigger alerts — handling gas sponsorship across supported chains so end users never need to hold native tokens. As of mid-2026, Compose had processed 2.8 million sponsored transactions on Alchemy's network with a 99.999% success rate.

Solana Support

Goldsky provides dedicated Solana integration through both Mirror and Turbo Pipelines. Developers can stream Solana blocks, transactions, instructions, and token activity — with full historical replay back to genesis — into the database destination of their choice. The platform delivers sub-second latency after confirmation and includes automatic reorg detection, ensuring applications always reflect canonical chain state. Both Mainnet and Devnet are supported.

Scale and Customers

The company reports supporting more than 155 blockchain chains, powering infrastructure that processes over $75 billion in total value locked. Customers include Uniswap, Kraken, Polymarket, and Phantom. Polymarket's real-time prediction market data runs on Goldsky's pipeline infrastructure, a demanding use case given the platform's need for low-latency event processing during high-activity periods.

The underlying architecture uses Apache Flink for stream processing and Redpanda for message queuing, enabling throughput at scale while maintaining data consistency. Goldsky achieved SOC 2 Type II certification in June 2026, covering security, availability, and confidentiality controls — a compliance milestone increasingly relevant as regulated financial firms adopt on-chain infrastructure.

Positioning

Goldsky occupies a layer that sits between raw blockchain nodes and application backends: it absorbs the complexity of indexing, reorg handling, schema management, and pipeline reliability so that product teams can focus on what they build rather than the plumbing underneath it. The company has positioned itself explicitly for "onchain finance" — applications that move money, settle trades, or track asset flows — rather than the broader web3 developer market, and the Compose product signals a push toward becoming an execution layer, not just a data layer, for that category.