> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/artifact-registry/3.0/get-started/migrate-to-harness-artifacts.md).

# Migrate to Harness Artifacts

Move your artifact infrastructure to Harness Artifacts, with concept mapping, migration paths, and what changes for your pipelines and clients.

Migrating an artifact registry changes how artifacts are stored, how dependencies are governed, how security findings are acted on, and how your AI-assisted delivery tooling reasons over what you ship.

This section covers the two most common migrations and helps you get your pipelines, clients, and teams pointing at Harness Artifacts with the least disruption possible.

## How a migration works

Every migration follows the same four steps, regardless of where you are coming from:

1. **Export or identify** the repositories and images you want to move.
2. **Create the destination registries** in Harness Artifacts: one registry per package type.
3. **Copy artifacts** using the Harness CLI migration tool (for supported sources) or Docker, skopeo, or crane.
4. **Re-point your clients and pipelines** at `pkg.harness.io` and verify that artifacts resolve correctly.

For a gradual cut-over with no downtime, you can use a Harness upstream proxy to serve traffic through the new registry while artifacts are still cached from your old source. Each migration guide covers both the bulk-copy and gradual approaches.

## How your concepts carry over

The following table maps what you already know to its Harness Artifacts equivalent. You will not be starting from scratch; the same ideas exist in a new shape.

| Concept                    | JFrog Artifactory                     | Amazon ECR                                               | Harness Artifacts                                                             |
| -------------------------- | ------------------------------------- | -------------------------------------------------------- | ----------------------------------------------------------------------------- |
| Where artifacts live       | Repository (local / remote / virtual) | Repository (per account, per region)                     | Registry                                                                      |
| Pull-through cache         | Remote repository                     | Pull-through cache                                       | Upstream proxy                                                                |
| Auth token                 | Access Token / Identity Token         | Identity and Access Management (IAM) + 12-hour ECR token | Personal access token (PAT) or Service access token (SAT) (no default expiry) |
| Vulnerability scanning     | Xray (separate product)               | Basic scanning + Amazon Inspector                        | Built-in scanning                                                             |
| Package policy enforcement | JFrog Curation (add-on)               | None                                                     | Dependency Firewall (OPA)                                                     |
| Lifecycle automation       | Retention policies                    | Lifecycle policies                                       | Cleanup and retention rules                                                   |
| Artifact signing           | GPG / RSA                             | Cosign (OCI)                                             | Signing keys (Debian, Alpine, Wolfi)                                          |
| Package formats            | 30+                                   | Docker and OCI only                                      | 25+                                                                           |

## What you gain by migrating

**A single platform for the whole delivery chain**. Harness Artifacts runs on the same platform as Harness CI, Harness CD, and your OPA governance policies. That means one RBAC model covers registries, pipelines, and deployments, with no separate license tier for scanning or curation.

**Shared policy enforcement across CI and Artifacts**. The same Open Policy Agent policies that gate your pipeline stages also govern what enters your registries through the Dependency Firewall. A package blocked by policy at the upstream proxy never reaches your build.

**Software Delivery Agent access to your artifact estate**. The Harness Software Delivery Agent can query your registries, artifact versions, scan results, lifecycle rules, and signing keys through the Artifacts API (the same API your pipelines use). The agent can answer questions like "what is the latest scan-clean version of this service?", surface which registries hold artifacts with open critical CVEs, or evaluate a lifecycle rule in dry-run mode before you apply it. Go to [Using Artifacts with AI agents](https://developer.harness.io/artifact-registry/3.0/agents/using-ar-with-ai-agents.md) for the full picture.

## Choose your migration path

{% content-ref url="/pages/xC9QWpSrawMYmlN2f8GO" %}
[Migrate from JFrog Artifactory](https://developer.harness.io/artifact-registry/3.0/get-started/migrate-to-harness-artifacts/from-jfrog-artifactory.md)
{% endcontent-ref %}

{% content-ref url="/pages/1VlMd2IX1KOguAgCrqmD" %}
[Migrate from Amazon ECR](https://developer.harness.io/artifact-registry/3.0/get-started/migrate-to-harness-artifacts/from-amazon-ecr.md)
{% endcontent-ref %}
