Stacks Datasets
Overview
Stacks Datasets are curated, public-good data products. Each dataset ships with a stable read API, a parquet download path, a schema reference, a dashboard view, and a freshness signal.
Datasets share the same publishing harness as Streams bulk dumps — parquet files partitioned by Stacks block height, refreshed on a schedule, with a machine-readable manifest at manifest/latest.json.
The shelf
| Dataset | Status | API | Parquet |
|---|---|---|---|
| STX Transfers Every canonical STX transfer event with sender, recipient, amount, and memo. | Shipped | /v1/datasets/stx-transfers | stacks-datasets/mainnet/v0/stx-transfers |
| PoX-4 / Stacking Solo, delegated, aggregated stacking calls — every PoX-4 contract call decoded with cycle math, BTC payout addresses, and signer keys. | Shipped | /v1/datasets/pox-4/calls | — |
| sBTC Deposits, withdrawals (create/accept/reject), signer-set rotations, governance hooks, plus SIP-010 mint/burn/transfer on sbtc-token. | Shipped | /v1/datasets/sbtc/events | stacks-datasets/mainnet/v0/sbtc |
| BNS BNS-V2 names, namespaces, marketplace listings, plus a current-state projection for fast resolve(fqn). | Shipped | /v1/datasets/bns/name-events | — |
| Network Health Daily rollup of canonical block count, average block time, and reorg count. | Shipped | /v1/datasets/network-health/summary | — |
How to read
Two access modes per dataset:
- Read API. JSON over HTTPS with cursor pagination. Best for dashboards, app event loops, and ad-hoc queries.
- Parquet download. Bulk historical files for warehouses, DuckDB, Spark, pandas. Best for ETL and analytics.
Read the manifest first; it points at every published file with row counts and SHA-256 checksums.
Pricing
Dataset reads are free for low-volume use. Heavy programmatic use rolls into Build or Scale via the standard API key path. Parquet downloads remain free regardless of tier.