Skip to content
For the complete documentation index, see llms.txt. This page is also available as Markdown.
WorkflowAdoptionIntegrationStandaloneHarness Platform

How Artifacts fits your workflow

Understand where Harness Artifacts sits in your delivery lifecycle, and how it works both inside the Harness platform and as a standalone registry.

Artifact Registry is the storage layer between your build tools and your deployment targets. Whatever you build and however you deploy, artifacts pass through a registry. Harness Artifacts is designed to fit that role without forcing you to change how your teams work, whether you use Harness end-to-end or bring your own CI, CD, and security tools.

This page maps out where Artifacts sits in a typical software delivery lifecycle and explains what the experience looks like for two adoption paths: teams running entirely in Harness and teams using Artifacts as a standalone service alongside other tools.

Where Artifacts sits in your delivery lifecycle

Every software delivery pipeline has a point where build outputs become versioned, stored artifacts that deployment pipelines consume. Artifacts is that point.

Stage
What Artifacts does

Build

Receives artifacts pushed from your CI pipeline (container images, packages, libraries, charts, or generic files) and indexes them with a version, digest, and timestamp.

Store

Stores artifacts in a globally replicated, deduplicated repository. Content is served from the region nearest to the consumer.

Proxy

Caches upstream open-source dependencies from public registries (Docker Hub, Maven Central, npm, PyPI, and others) so your builds do not depend on external availability.

Scan

For Docker and Helm registries, automatically generates an SBOM and scans every pushed image for vulnerabilities. For upstream dependencies, Dependency Firewall evaluates inbound packages against your security policies before they are cached.

Deploy

Serves artifacts to deployment pipelines, Kubernetes clusters, ECS tasks, Lambda functions, and any consumer that speaks the standard protocol for the format.

Govern

Enforces RBAC, lifecycle cleanup rules, retention rules, and OPA policies. Records every push and pull in the audit trail.

Artifacts is not just a storage bucket. Every artifact that flows through it is scanned, versioned, governed, and auditable, without any extra tooling or pipeline steps.


Use case 1: Artifacts within the Harness platform

If you use Harness CI and CD, Artifacts is the native artifact layer, already wired in. No connectors to configure, no separate authentication setup, no external credentials to manage. The pipeline identity that runs a stage is automatically authorised to push or pull from any registry the identity has access to.

In Harness CI

When you build in Harness CI, Artifacts is the default registry type for artifact steps:

Build and Push to Docker: Select a Docker registry in the step configuration. Harness CI authenticates automatically using the pipeline identity. No Docker Hub credential or external connector required.

Upload Artifacts to Harness AR: A dedicated step for non-container formats. Choose the registry, set the package type, point to the build output path. The step uploads the artifact and creates the version entry in the registry. Supports Maven, npm, Python, NuGet, Conda, Generic, and other formats.

Run and Plugin steps: Reference a container image stored in Artifacts directly from a Run or Plugin step. Select the registry in the step configuration and enter the image reference. No connector needed.

The result: developers define the artifact destination at the step level, not at the pipeline or account level. The same registry list that appears in the step selector is what the team created in the Registries view: one consistent, permission-controlled list.

In Harness CD

CD services reference Artifacts as a native artifact source. In the service configuration, select Harness Artifacts as the source and choose the registry and artifact. During pipeline execution, Harness fetches the artifact automatically using the pipeline identity.

There is no connector to create, no username or password to provide, and no credential rotation to maintain. The deployment stage sees the same artifact list the build stage produced. Traceability from build to deployment is built in: the Deployments tab on the artifact version detail page shows every environment the artifact has been deployed to.

In Harness STO and SCS

Security scanning integrates directly with Artifacts through the HARNESS ARTIFACT SCAN PIPELINE, which runs automatically when you push to a Docker or Helm registry that has scanning enabled. The pipeline:

  • Generates an SBOM (via Harness Supply Chain Security) and attaches it to the artifact.

  • Scans the image for vulnerabilities using Aqua Trivy or Grype (via Harness Security Testing Orchestration).

  • Evaluates the results against any Security Tests policy set attached to the registry.

  • Quarantines the image automatically if it violates the policy.

No connector configuration is needed to reference a Harness Artifacts-hosted artifact from an SCS or STO pipeline step. Results appear on the SBOM and Vulnerabilities tabs of the artifact version detail page.

Governance

Policy enforcement uses the same OPA-based Policy as Code model used across the Harness platform. You can write policies that gate artifact creation, registry configuration changes, and lifecycle rule execution. Policies evaluate automatically; no manual approval gates are required unless you configure them.

Every artifact push and pull, every registry create and delete, and every lifecycle rule execution is recorded in the Audit Trail and the Artifact Registry Audit Dashboard.

No connector setup

The native integration between Artifacts and the Harness platform means pipeline identities (service accounts, pipelines, and users) authenticate to Artifacts using their Harness identity. No Docker Hub credentials, Maven settings files, or npm .npmrc files with passwords are needed inside Harness CI/CD pipelines. Standard client credentials are only required when pushing or pulling from outside the Harness platform.


Use case 2: Artifacts as a standalone registry

You do not need to use Harness CI or CD to use Artifact Registry. Artifacts speaks the native protocol for every format it supports. If your existing tools already push to Docker Hub, npm, PyPI, Maven Central, or another standard registry, they can push to Artifacts with a credential swap: no plugin, no SDK, no migration tooling required.

With any CI system

Jenkins: Install the Harness plugin to authenticate the Harness CLI (hc) in your pipeline, then use standard client commands (docker push, mvn deploy, npm publish) pointed at pkg.harness.io. For artifacts that do not have a standard push command, use hc artifact push.

GitHub Actions: Use the Setup Harness CLI action to install and authenticate hc, then push with your standard client. The CLI handles authentication so your workflow does not store registry passwords in secrets.

Any other CI: Install the Harness CLI in your build environment, authenticate with an API key, and push using standard clients. Artifacts behaves as a drop-in replacement for any registry that speaks the same protocol. A Maven build configured for Maven Central points at pkg.harness.io instead; the mvn deploy command is unchanged.

With standard clients

Artifacts does not require Harness-specific tooling. Every format uses its standard client:

Format
Push command
What changes

Docker

docker push pkg.harness.io/<account>/<registry>/<image>:<tag>

Registry host and path only

npm

npm publish --registry https://pkg.harness.io/...

Registry URL in .npmrc

Maven

mvn deploy

Repository URL in pom.xml or settings.xml

Python

twine upload or uv publish

Index URL in .pypirc or environment variable

Cargo

cargo publish --registry harness-<registry>

Registry in ~/.cargo/config.toml

Helm OCI

helm push chart.tgz oci://pkg.harness.io/...

Registry host only

The authentication token is a Harness API key used as the password. Your username is your Harness account email. The Setup Client panel on any registry walks you through the exact commands for the format you are using.

With the Harness CLI

The hc CLI is the command-line interface for Artifacts. It covers operations that standard clients do not: searching for artifacts, copying versions between registries, uploading generic files, evaluating Dependency Firewall policies from a script, and managing registry metadata.

Install hc from Harness CLI. Go to CLI reference for the full command reference.

With the REST API

The Artifacts REST API covers every operation available in the UI: creating registries, listing artifacts, retrieving version metadata, triggering lifecycle rules, and managing policy exceptions. The API base is https://app.harness.io/har/api/v1/. Authentication uses x-api-key: <your-api-key>.

Go to REST API for the full reference.

With Terraform

The Harness Terraform provider includes resources for creating and configuring Artifacts registries, upstream proxies, and lifecycle rules. You can manage your entire registry structure as code and apply it through your existing infrastructure pipeline.

Go to Terraform provider for the resource reference.


What both paths share

Regardless of whether you use Artifacts inside the Harness platform or as a standalone service, the following capabilities apply to every registry and every format:

RBAC: Three built-in roles (Viewer, Contributor, Admin) apply at the registry level. You control who can push, pull, and manage each registry using Harness RBAC, which integrates with your existing SSO and identity provider.

Upstream proxies and Dependency Firewall: Every format supports upstream proxy registries. Attach Dependency Firewall policies to evaluate inbound open-source packages for vulnerabilities, license violations, malicious packages, and package age before they enter your registry. Go to Dependency Firewall.

Lifecycle rules: Cleanup and retention rules apply to all formats. Schedule automatic cleanup based on artifact age, last download, or version count. Go to Lifecycle Management.

Multi-region replication: Artifacts are stored in your primary region and replicated globally. Pulls are served from the nearest regional cache. Supported regions: US-West, US-East, EU-West, EU-East, Asia-Pacific, and Oceania.

Audit trail: Every push and pull is recorded in the Artifact Registry Audit Dashboard with username, client IP, and timestamp. Registry configuration changes are recorded in the Harness platform Audit Trail. Go to Artifact Registry audit dashboard.

Storage-based pricing: You pay for the storage you use. Harness deduplicates content at the account level, so pushing the same layer twice does not double your storage.


Next steps

Last updated

Was this helpful?