Safeguard
Supply Chain•

Which databases feed OSV, and which are converted from other formats?

OSV combines native format advisory databases with converted Debian, Alpine and NVD data. See the source list, what conversion changes and why coverage varies.

Safeguard Security Team
3 min read

OSV.dev aggregates two kinds of sources. Native sources publish in the OSV format themselves, including the GitHub, PyPI, Go and Rust advisory databases. Converted sources, which OSV.dev's pipeline transforms, are Debian Security Advisories, Alpine SecDB and NVD CVEs for open source software.

The source list

From the OSV data page, fetched 19 October 2026:

TypeSources
Native OSV formatGitHub Advisory Database, PyPI Advisory Database, Go Vulnerability Database, Rust Advisory Database, Drupal Advisory Database, Global Security Database, OSS-Fuzz, Rocky Linux, AlmaLinux, Haskell Security Advisories, RConsortium Advisory Database, OpenSSF Malicious Packages, Python Software Foundation Database, Bitnami Vulnerability Database, Ubuntu, opam, and the Erlang Ecosystem Foundation CNA
Converted by the pipelineDebian Security Advisories, Alpine SecDB, NVD CVEs for open source software

Why the split matters

A native source controls its own records, including ranges and aliases. A converted record is a translation. The OSV FAQ explains why OSV exists: CPEs are hard to match automatically to package names and versions. A conversion from CPE style data has to infer package and range, and inference can be wrong or broad.

So when two scanners disagree on a flaw in a language library, ask which source each used. A native ecosystem record with precise ranges and a converted record from a general feed can give different answers. For background see vulnerability tool advisory data sources.

Ecosystems covered

The same page lists covered ecosystems, including crates.io, Go, Maven, npm, NuGet, Packagist, PyPI, RubyGems, Hex, Pub, GitHub Actions, Debian, Alpine, Ubuntu, the Linux kernel and others. Coverage of a registry is only as good as the sources behind it. An ecosystem on the list can still lack advisories if no source publishes them.

How to use this in practice

  1. Take one dependency you care about.
  2. Search its name in OSV and note each record's ID prefix, which hints at the source.
  3. Check whether a native source record exists beside a converted one.
  4. Compare ranges. Prefer the record with explicit fixed events.
  5. If a record is missing, the FAQ says to raise it with the home database that owns it. OSV may have rejected the import on quality grounds, and the not found page for that record says so.

Limits of this list

The list changes. New sources are added, and older ones can be removed. Treat any copy of it, including this one, as dated. Re-read the page before you rely on it in a vendor review or audit answer.

Try it yourself

Click through it at your own pace, or open it full screen.

Look at which ecosystems appear in the inventory, since each one depends on a different source set.

Safeguard SBOM and dependency scanning covers components across ecosystems, with reachability analysis on top. A free scan shows your own mix.

Quick check

Is a converted record less reliable?

Not always, but conversion involves inference. Check ranges against the vendor advisory.

What if OSV has no record for a known flaw?

It may have failed a quality check on import. The FAQ advises raising it with the database that owns the record.

Sources

Last reviewed: 19 October 2026.

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.

Self-healing security runs on Safeguard.

Your first fix PR is minutes away.

No sales call required, even your agent can complete the purchase over MCP.