chainstar
Raw blockchain data to clean dimensional tables in your own warehouse, from one command. In development. Python, Apache-2.0. Runs on Linux, macOS, Windows.
What it does
chainstar turns raw blockchain data into clean, analysis-ready tables in a warehouse you own. One command runs the whole pipeline:
- Extract blocks, transactions and logs from the chain.
- Decode them into readable events.
- Load the result to Parquet.
- Model it into a star schema with dbt.
You bring your own RPC URL. Nothing passes through a Bit Primitive server.
The tables
| Facts | Dimensions |
|---|---|
| Transfers | Wallets |
| Swaps | Tokens |
| Fees | Contracts |
| Dates |
How it is built
- Extraction is built on cryo.
- Modelling is a dbt package, so the tables can be extended, tested and documented the same way as the rest of your dbt project.
First release
- Chains: Ethereum and Base
- ERC-20 transfers decoded
- Warehouse targets: DuckDB and Postgres
Roadmap
- More EVM chains, then Bitcoin and Solana
- BigQuery, Snowflake and other warehouses through dbt adapters
- Homebrew and Docker distribution