AI-Native Databases
Built in Europe
High-performance vector and document databases built in Erlang. Open source, embeddable, production-ready.
Two Databases, One Platform
Purpose-built databases for AI workloads. Use them together or separately.
Barrel Vector
Apache 2.0Vector Database
High-performance single-node vector database with HNSW, FAISS, DiskANN, and BM25 backends. Perfect for semantic search and RAG applications.
- + <5ms search latency
- + Hybrid search (vector + BM25)
- + Metadata filtering
- + Embeddable in Erlang/Elixir
Barrel Docs
Apache 2.0Document Database
Document database with MVCC, declarative queries, real-time subscriptions, and peer-to-peer replication. Built for edge computing and sync.
- + <1ms read latency
- + P2P replication
- + Real-time subscriptions
- + Automatic path indexing
Why Barrel Databases?
Built different. Built in Europe. Built to last.
Open Source
Apache 2.0 licensed. Fork it, modify it, deploy it anywhere.
Built in Erlang
Battle-tested VM. Fault-tolerant by design. No garbage collection pauses.
Made in Europe
European company. No US Cloud Act. GDPR compliant by design.
Embeddable
Use as a library in your Erlang/Elixir apps. No separate process needed.
Powers Barrel Platform
These databases are the foundation of Barrel Platform — privacy-first AI agent networking for Europe. The same battle-tested code you can run on your own infrastructure.
Production Tested
Powers real services
Privacy First
GDPR by design
100% Open Source
All components are Apache 2.0 licensed. Fork it, modify it, deploy it anywhere. No vendor lock-in, no usage fees.
Ready to Get Started?
Self-host free with our open source databases, or contact us for managed hosting.