Tiders
Open-source Python framework for continuous onchain data pipelines with provider portability and Rust-powered decoding.
Best for: Data engineers who want a self-hosted, continuously synced onchain data warehouse without writing chain-head polling, decoding, or retry logic from scratch.Tiders is an open-source Python framework for building continuous onchain data pipelines. A pipeline is composed of four parts: a Provider (HyperSync, SQD, or a raw RPC), a Query that defines which events and traces to fetch, Steps that transform the data, and a Writer that persists results to a target store.
Decoding and transformation are handled in Rust under the hood, which keeps throughput high even for large historical backfills. The output plugs into Polars, Pandas, or DataFusion for in-process analytics, and Tiders includes Writers for ClickHouse, DuckDB, Postgres, Iceberg, Delta Lake, and Parquet out of the box.
The key design principle is provider portability: switching from HyperSync to SQD or to a raw RPC requires changing one line. This insulates your pipeline from provider outages or pricing changes without requiring a rewrite.
Tiders was forked from steelcake/cherry (the original SteelCake team, who built HyperSync at Envio). The project is at v0.1.8 and is under active development. A companion project, tiders-x402-server, exposes the synced warehouse as an agent-payable SQL endpoint using the x402 payment protocol.
Pricing
Open-source and free (MIT/Apache-2.0). Self-hosted: you pay for your own infrastructure.
Chains Supported
EVM chains via HyperSync and SQD; any RPC-compatible chain via the RPC provider
- Continuously syncing decoded onchain events into ClickHouse, DuckDB, or Postgres
- Building multi-chain data pipelines with a single Python codebase
- Historical backfills with Rust-powered decoding for high throughput
- Owning your onchain data warehouse without a managed indexing vendor
- Exposing a synced warehouse as an SQL endpoint via the x402 payment protocol
