vedro

vedro is a simple, self-contained S3-like object storage written in Go. Built for local use — no cloud dependencies, no databases required.

⚠️ Not recommended for large-scale production, but works great for:

• Local artifact storage
• Simple CI/CD pipelines
• Debugging and local development
• Temporary S3 stubs

Features

• S3-compatible interface (ListBucket, GetObject)
• ETag, Last-Modified, and If-Modified-Since support
• S3-standard XML responses
• Lightweight HTTP server
• Path traversal protection
• Logging and panic recovery middleware

Configuration

Defaults are defined in config/config.go and can be overridden via CLI flags at runtime:

Example:

./vedrod -addr :9000 -root /data/buckets -log=false

Usage

1. Build

go build -o vedrod .

2. Prepare directory structure

/var/vedra/
├── bucket1/
│ ├── file1.txt
│ └── image.png
└── bucket2/
 └── report.pdf

3. Run

./vedrod

Example requests

List bucket objects (S3 ListBucket)

GET http://localhost:8080/bucket1

Response (XML):

<ListBucketResult>
 <Name>bucket1</Name>
 <Contents>
 <Key>file1.txt</Key>
 <LastModified>2025年04月18日T17:42:18Z</LastModified>
 <Size>123</Size>
 <ETag>...</ETag>
 </Contents>
</ListBucketResult>

Download an object (S3 GetObject)

GET http://localhost:8080/bucket1/file1.txt

Notes

• Does not support Multipart Upload, ACL, PUT/DELETE, presigned URLs, etc.
• Assumes all data is available locally on disk.
• MIME types are detected automatically via http.DetectContentType.
• No caching — reads directly from disk on every request.

License

GNU GPLv3