A purpose-built Rust engine.
A storage engine built from scratch, with C++ AVX-512 extensions for vector math. No SQLite. No RocksDB.
Explore the architectureDocuments, relationships, and semantic search.
One database. A more natural way to build.

Purpose-built from the storage up.
Relationships belong with your data.
Find meaning, beyond exact matches.
A more natural way to work with data.
Flexible like documents. Connected like a graph. Thoughtful enough to search by meaning. It all comes together in CleaveDB.
Store flexible JSON in named buckets. No rigid schema, no setup step. Just POUR your first document.
// A document. A bucket. That's it.
POUR INTO users "jane" {
"name": "Jane",
"age": 25,
"city": "Manila"
}Less time stitching infrastructure together.
More time building what comes next.
A storage engine built from scratch, with C++ AVX-512 extensions for vector math. No SQLite. No RocksDB.
Explore the architectureA Go-based coordinator connects to the Rust engine for distributed, multi-shard deployments.
Learn about scalingAuthentication, tenant isolation, and granular document-level security keep access rules close to your data.
Understand securityWe believe building scalable architecture shouldn’t require years of specialized SQL knowledge. Our mission is to make high-performance data storage inherently readable, learnable, and accessible to developers of any background through simple, conversational semantics.
A future where backend complexity simply gets out of the way. We envision a world where bridging complex graph relationships and neural AI search is as easy as storing a JSON document, empowering you to build the next generation of software without limits.
“Data infrastructure shouldn’t be a barrier to entry. It should be the invisible bridge between a great idea and its execution.”
John David L. Perez
Your next idea starts with your first query.