// the find
subsquid/squid-sdk
TypeScript ETL toolkit for indexing Ethereum, Solana, and Substrate data, sourced from SQD Network.
Squid SDK is a TypeScript ETL framework for pulling blockchain data — Ethereum, Solana, Substrate, and now Bitcoin — into your own storage. Unlike TheGraph or Ponder, it's a toolkit: you write the transformation logic, point it at a data source (either SQD Network's decentralized data lake or a plain RPC), and pipe the output wherever you want. The target audience is teams who need custom indexing logic without being locked into a managed platform's query model.
The binary codecs (evm-codec, SCALE for Substrate, Borsh for Solana) are genuinely fast and type-safe — you get decoded, typed data rather than raw hex blobs. The dual data-source design (SQD Network archive for historical bulk sync, RPC for real-time tip-following) is a practical architecture that handles the most painful part of blockchain indexing: catching up from genesis without hammering an RPC. The monorepo is well-structured with Rush, per-package changelogs, and fixture-based tests against real block data, which is the right way to catch normalization bugs. Active maintenance is visible — evm-rpc and evm-normalization had commits today.
The dependency on SQD Network for fast historical sync is a meaningful operational bet: if the network has gaps or goes down, you fall back to slow RPC crawling with no graceful degradation story documented. The README drops you into a CLI scaffold without explaining what the generated code actually does, so the learning curve for anyone not already familiar with the TheGraph mental model is steep. Bitcoin support looks early — it's a separate directory tree with its own data service rather than being a first-class chain in the processor abstraction. Error handling and backpressure strategies for long-running processors aren't documented, which matters a lot when you're indexing 20 million blocks and the RPC starts rate-limiting you.