Open source · 25k+ GitHub Stars

Simple, scalable,
reliable storage

A distributed store built on O(1) disk reads and append-only writes — data lands fast and comes back fast at any scale. One binary serving S3 objects, Iceberg tables, and POSIX files.

Free
Dev & Test under 25TB
$2/TB
Per Month
$20/TB
Per Year
Trusted by
Features

One binary. Objects, tables, and files.

SeaweedFS is a simple and highly scalable distributed file system, built to store billions of files and serve them fast! In the spirit of Google File System, a single distributed engine serves as your S3 object store, Iceberg table warehouse, and POSIX filesystem — all from the same cluster.

Object Store

S3-compatible distributed object store. Scales linearly from gigabytes to petabytes with sub-millisecond lookups.

Iceberg Catalog & Maintenance

A built-in Apache Iceberg REST catalog over S3 table buckets — no separate catalog service to deploy. It also maintains your tables automatically, compacting Parquet files and cleaning up in the background. Query from Spark, Trino, Dremio, DuckDB, and more.

FUSE Mount

100% POSIX compliant — hardlinks, permissions, and atomic renames all work. Peer-to-peer chunk sharing between clients accelerates repeated reads across the fleet.

More capabilities

O(1) Disk Reads

Every file resolves to a single disk seek, so read latency stays flat and sub-millisecond whether you store millions or billions of files.

Append-Only Writes

Writes only ever append — never overwrite in place. Sequential I/O keeps ingestion fast, avoids fragmentation, and makes data immutable and recoverable by design.

Linear Scalability

Add nodes to scale capacity and throughput predictably. Proven at petabyte scale with billions of files.

Data Recovery

Restore deleted files within a configurable retention window. SeaweedFS delays garbage collection and replays the filer metadata log so operators can undelete from the Admin UI — no snapshot restore required.

Point-in-Time Recovery

Roll a folder, bucket, or object back to exactly how it looked at a chosen moment. PITR reverts overwrites and restores deletes across a whole subtree at once, previewing every change before it touches your data.

Self-Healing Storage

Automatically detects and removes corrupted data entries, protecting your data even after unexpected server shutdowns.

Customizable EC Ratio

Optimize storage costs with flexible erasure coding. Configure ratios like 20+4 for just 1.2x storage overhead.

Automatic EC Repair

Detects and rebuilds missing or corrupted erasure coding shards automatically, ensuring continuous data availability.

Iceberg Table Maintenance

The catalog keeps tables fast and lean on its own — compacting small Parquet data files, rewriting delete files, expiring old snapshots, removing orphaned files, and rewriting fragmented manifests as tables grow.

Admin UI with OIDC

Let operators sign in through your existing identity provider with mapped admin or read-only access controls.

zstd Storage Compression

New data is compressed with zstd by default — typically 10–30% smaller than gzip and several times faster to decode, with no format change and nothing to migrate.

Transparent Pricing

Simple, predictable pricing: $2/TB/month or $20/TB/year. Free for dev & test under 25TB. No hidden fees or complex tiers.

HDFS & Kubernetes

Works as a drop-in HDFS replacement and runs natively on Kubernetes with a first-class Helm chart and CSI driver.

How It Works

Deploy in Minutes, Run for Years

SeaweedFS Enterprise extends the battle-tested open-source SeaweedFS with enterprise data protection. Your existing setup stays the same — just add a license key.

  • Same deployment, same APIs, same tools
  • Self-healing runs in the background automatically
  • No vendor lock-in — your data stays portable
  • Free for dev & test under 25TB with no feature gates
Deploy Now
# Start with your existing SeaweedFS cluster
$ weed server -master.port=9333
 
# Apply your enterprise license
$ weed shell
> license.set -file=license.key
 
# Self-healing is now active
✓ License verified
✓ Self-healing enabled
✓ EC repair active
Comparison

Open Source vs. Enterprise

Enterprise adds critical data protection features on top of the open-source foundation you already trust.

Feature Open Source Enterprise
S3 / FUSE / HDFS API
Erasure Coding Fixed ratio Customizable ratio
Data Recovery
Point-in-Time Recovery
Self-Healing Storage
Automatic EC Repair
Automatic EC Vacuum
Admin UI with OIDC
Storage Compression gzip zstd by default + gzip
Priority Support

Ready to Protect Your Data?

Evaluate up to 25TB free for development and testing. No credit card required. Deploy self-healing storage in minutes.