Open data
Data dumps
PkgSeek's public package and CVE datasets, free to download and build on — no signup, no rate limit, no 10,000-row cap. Columnar Parquet for analytics and newline-delimited JSON for everything else. Regenerated on each indexing run.
Last generated
packages
One row per package build from a public distribution's official repository, including repository provenance, version, architecture, sizes, and package metadata.
cves
Package-to-CVE links for public official repositories, including repository and advisory provenance plus CVSS scores when available.
Query it in one line
Parquet loads straight into DuckDB, pandas, Polars or Spark — no import step.
-- DuckDB: newest openssl across every distro
SELECT distro, version FROM 'packages.parquet'
WHERE name = 'openssl' ORDER BY distro;
# pandas
import pandas as pd
df = pd.read_parquet('packages.parquet')