Alarik

Alarik is a self-hosted Amazon S3 alternative built for teams that want full control over their object storage without depending on managed cloud infrastructure or restrictive licenses. It's written in Swift and SwiftNIO, which gives it low-latency I/O, highly parallel request handling, and no garbage collection pauses. Apache 2.0 licensed, no strings attached.

The S3 compatibility is genuine. Point your existing AWS SDK, rclone, or s3cmd at an Alarik endpoint and it works. No code changes, no adapter layers.

Key capabilities:

• S3-compatible API Drop-in replacement for existing S3 workflows and tooling
• Built-in web console A Nuxt-based UI for managing buckets and objects without touching a CLI
• Single-node to cluster Start with one container; scale to a Postgres-backed, sharded, erasure-coded cluster only when you need to
• Rendezvous hashing Distributes object data across nodes predictably in cluster mode
• Reed-Solomon erasure coding Protects data across shards without full replication overhead
• Memory-safe architecture Swift's ARC model eliminates buffer overflows and use-after-free bugs

Deployment starts with Docker Compose: two containers, one command. There's no configuration sprawl to manage before you can start storing objects. For teams migrating away from MinIO or looking for a lighter alternative to Ceph, Alarik positions itself as actively maintained, genuinely open, and architecturally simpler to start with.

It fits local development environments, application backends handling user uploads and media, backup workflows, and self-hosted replacements for Cloudflare R2 or similar cloud storage. The opt-in scaling model means you're not forced to operate a distributed system until the workload actually demands it.