Company
Selling data the way infrastructure deserves
datastore.sh exists because every serious trading and research team was building the same blockchain data pipelines, badly duplicated, at enormous cost. We build them once, to an auditable standard, and sell the output.
Operating principles
Four rules we don't break
01
Depth over breadth
Two chains covered completely beat ten chains covered approximately. We add coverage when we can hold it to the same standard as what we already sell — and not before.
02
Files over access
We sell Parquet you own, not API quotas you rent. Your queries run on your infrastructure, your costs are your own, and losing a vendor never means losing your data.
03
Evidence over assurance
Checksums, manifests, lineage records, and immutable versions — every claim about a dataset is machine-checkable. Trust should not require believing us.
04
Corrections in the open
Chain data pipelines produce defects; pretending otherwise is the real red flag. When we find one, it becomes a new version with a change log, delivered free to affected buyers.
What we operate
Chain capture infrastructure for Solana and Hyperliquid, protocol decoders pinned to program versions, a continuous reconciliation system, and the publication pipeline that turns all of it into versioned Parquet datasets. The methodology is documented publicly on the platform page and in our engineering notes.
Reach us
Sales, coverage requests, procurement, or a technical question about a dataset — it all lands with the team that builds the pipelines.