Overview
Introduction
What Relic is, what it serves, and the parts an installation runs.
Relic is a self-hosted artifact repository in one binary. It keeps content in S3-compatible storage, Azure Blob Storage, Google Cloud Storage or a local directory, and metadata in PostgreSQL or SQLite.
Manage it with the relic CLI, the web UI or the Relic API.
What it serves
| Area | Contents |
|---|---|
| Repositories | npm, Maven and Gradle, Docker and OCI, Go, PyPI, Helm, NuGet, RubyGems, Cargo, Conan, Terraform, Hugging Face, Nix, Debian, RPM, Alpine, generic files, and MCP servers and agent skills; local, remote (caching proxy) or virtual. |
| Builds, release bundles, evidence | What was built from what, where it was promoted, who signed it. |
| Relic Scan | Packages, licenses and vulnerabilities (Syft, Grype), malicious packages, secrets (Gitleaks), misconfigurations (Trivy). Policies can block downloads. |
Architecture
| Part | Role |
|---|---|
relic server start | The HTTP APIs and the web UI. |
relic server worker | Background jobs: scanning, garbage collection, replication, webhooks, cleanup, retention. relic server start --worker runs both in one process. |
| PostgreSQL or SQLite | Metadata. SQLite is for a single node. |
| Object storage | Content, stored once per SHA-256. See Storage. |
Next steps
- Run a server: Installation.
- Use an existing server: install the relic CLI, then sign in.