your #RAG stack doesn’t need a database server. On Tuesday we opened OneBucket™
By Shayne Stubbs · November 13, 2026 · Curated by George's Blog
your #RAG stack doesn’t need a database server.
On Tuesday we opened OneBucket™ to developers. Today I want to highlight a project by Alessandro Pireno. His distributed object store RAG is a complete pipeline you can read in an afternoon at https://lnkd.in/eHZ7V2u9
Drop a file in a bucket. It gets chunked, embedded and indexed. Then you query it from Claude over MCP.
The interesting part is what’s missing:
→ no database server: the index lives as files in object storage
→ no vector database service to run or keep in sync
→ no code change to move: one environment variable switches it from Amazon Web Services (AWS) S3 to s3.OneBucket.io
On #OneBucket, the index is written to every global region. The documents themselves fill in only where they’re actually read.
Your data stays where it lives. Your agents can still find it anywhere.
It’s open source (MIT). Fork it, point it at your own buckets, and tell us what you build.
Thank you, Alessandro Pireno
Start building at s3.onebucket.io: 90 days free if you sign up by December 31.
🔗 https://lnkd.in/e6cBXb7f
🔗 onebucket.io/developers
#RAG #AIagents #developers #MCP #OneBucket