the product

One coordinate in front of every package manager you already run — then identity, provenance, advisories, an SBOM, and a full audit trail attached to each artifact. Point your existing clients at it; nothing new to learn.

See it in action

Real screens from the registry: the whole supply chain at a glance, a malicious npm package caught at intake, and exactly who pulled it.

Dashboard — inventory & findings at a glance

Caught — a malicious npm package & its blast radius

Audit — who pulled what, when, and from where

The gastropod dashboard — counts for packages, versions, artifacts, pulls, consumers and repositories; a security band flagging advisories, identity and provenance mismatches, and verified/attested artifacts; plus pulls-over-time and most-pulled-packages charts. — One view — inventory, advisories, identity & provenance flags, verified/attested counts, and pull volume.

The package page for the npm package event-stream, flagged with an identity mismatch — it declares itself as a different name than it was served under — showing its transitive dependents (the org's own services that pulled it) as the blast radius, and the full list of consumers, each with a one-click CycloneDX SBOM. — A malicious npm package, caught: `event-stream` flagged on an **identity mismatch** — it declares a different name than it was served under — with the blast radius (the services that pulled it) and every consumer, each with a CycloneDX SBOM.

The audit log — a stream of per-pull events, each with a timestamp, the package PURL (including the flagged event-stream and crossenv packages), the principal who pulled it, the client, the source IP, and the HTTP status. — The audit spine — every pull attributed to a principal: who, what, from where, and when.

Every ecosystem, one registry

Hosted or proxied across Go modules, npm, Debian/Ubuntu, OCI containers, Python (PyPI), Maven & Gradle, Alpine (apk), RHEL family (rpm), and .NET (NuGet), plus on-demand source proxies for GitHub and GitLab. Set one variable and your toolchain just works.

Go modules

GOPROXY=…/go

npm

registry=…/npm/

Debian / Ubuntu

deb …/deb/<distro>

OCI containers

docker pull …/<repo>

Python (PyPI)

--index-url …/pypi/

Maven & Gradle

repository …/maven/

Alpine (apk)

…/apk/<repo>

RHEL family (rpm)

baseurl=…/rpm/

.NET (NuGet)

nuget source …/nuget/

…more coming soon.

Connect in one line

Point your existing clients at gastropod — no new tooling. Replace TOKEN with the value shown when you issue one; <org>/<repo> is a hosted repository.

point a client at it

# credentials go inline in GOPROXY
export GOPROXY=https://you:TOKEN@gastropod.io/go
export GOSUMDB=off   # private modules aren't in the public checksum db

How it works

A package flows through gastropod on its way to a build, and picks up identity, findings, and provenance it carries everywhere that exact digest appears.

See

An artifact appears via a proxy pull, a push, or a CI job.
Resolve

Identity is fixed: PURL, version, ecosystem, source, and sha256 digest.
Verify

Checksums and provenance are checked — Go's sum.golang.org h1, npm SRI, Sigstore where present — and OSV advisories correlated.
Attribute

Every pull is recorded to a principal in the access spine: who, what, from where, when.
Remember

Advisories, the CycloneDX SBOM, provenance, and dependents attach to the artifact identity.
Reuse

The next appearance of that exact digest inherits the intelligence — no rework, across your org.

gastropod vs. a traditional artifact repository

Artifact repositories move bytes well — gastropod does that too, and folds in the intelligence and audit layer they leave to add-ons and higher tiers.

Capability	gastropod	Traditional artifact repo
Multi-ecosystem proxy & hosted registry	built in	built in
Content-addressed, deduplicated storage	core design	partial
PURL-native artifact identity	first-class	partial
Provenance verification (Go h1, npm SRI, Sigstore)	built in	add-on
CycloneDX SBOM on demand	built in	add-on
OSV advisory correlation at the artifact	built in	separate product
Blast-radius / transitive-dependent graph	built in	limited
Per-pull attribution audit spine	core design	higher tier
Scanner findings ingest (bring your own)	built in	own scanner
Self-hostable	yes	yes

built in / first-class available, often an add-on or higher tier limited or not a focus

A directional map of how gastropod sees the category — security and intelligence features are generally separate products or higher tiers in traditional artifact repositories. Verify specifics against each vendor's current documentation.

The full feature list

Every capability is on every tier — what changes between tiers is rates, limits, and SLA, not which features you're allowed to use.

registry & ecosystems

Proxy & hosted repos for Go, npm, Debian/Ubuntu, OCI, PyPI, Maven & Gradle, Alpine apk, RHEL rpm, NuGet
On-demand source proxies for GitHub & GitLab
Content-addressed sha256 storage, deduplicated across the instance
Virtual / aggregate repositories
Point existing clients at it — no extra tooling

artifact intelligence

PURL-keyed identity and where-seen history
OSV advisory correlation per package version
CycloneDX SBOM reconstructed on demand
Direct & transitive blast-radius graph
Provenance: Go sum.golang.org h1, npm SRI, Sigstore, signed apt
Identity- and provenance-mismatch flags

control & audit

Per-pull access spine — every pull attributed to a principal
Exportable audit log and access events
Block rules at the point of intake
Scanner findings ingest (Trivy, Grype, Snyk, Anchore; CycloneDX, OpenVEX, JSON)

access & security

SSO via OIDC — on every tier
Role-based access control (admin / publisher / reader)
Scoped API tokens
Encrypted in transit and at rest
Per-tenant isolation

deployment & support

Fully-managed cloud, or self-hosted
Air-gapped deployment (Enterprise)
99.9% uptime SLA (Business); custom SLA (Enterprise)
Full audit log and access events

See it on your own artifacts.

See plans →

get on the waiting list book a demo