# Hanzo S3 — Hanzo AI

> Object storage that your existing tools already speak, and you never pay to read your own data back.

Storage

# Hanzo Object Storage

Buckets and objects, presigned links, multipart uploads, versioning and object lock.

[API reference](https://docs.hanzo.ai/docs/openapi/s3)[Source](https://github.com/hanzoai/s3)

An object, kept

- One binary, four roles01
- Keys from KMS02
- Ten data shards, four parity03
- One protocol04

One binary in four roles: a master that tracks where everything is, volume servers that hold the bytes, a filer for the namespace and the HTTP gateway your code talks to.

Server-side encryption whose keys are minted by Hanzo KMS rather than read out of a config file.

Objects are erasure-coded across ten data shards and four parity shards, so four can be lost and the object still reads.

Every call between the pieces speaks one binary protocol, with no gRPC anywhere: one transport to secure and one to debug.

Paste and ship

Paste this into any agent. It reads the skill manifest and calls Object Storage from there.

Agent promptapi /v1/s3/bucketsCopy

```
Read https://hanzo.ai/skill.md and use Hanzo Object Storage in my project. Start with: hanzo s3 buckets list
```

## Get Object Storage

Distributed object storage

License: Apache-2.0[hanzoai/s3](https://github.com/hanzoai/s3)

[Deploy to Cloud](https://platform.hanzo.ai/deploy)