Nexus DB v2.0 is officially active
Build with the speed of light.
A globally distributed database built specifically for vector embeddings, time-series operations, and sub-millisecond transactions.
nexus-shell - v2.1-cli
CHOOSE QUERY PROTOCOL
nexus://admin >
SELECT id, name, cosine_similarity(embedding, [0.15, -0.42, 0.89]) AS sim
FROM products
ORDER BY sim DESC LIMIT 2;system_output //
Query completed in 1.48 milliseconds.
[
{ "id": "p_8902", "name": "Titanium Smart Ring", "sim": 0.9421 },
{ "id": "p_1143", "name": "Carbon Chrono Watch", "sim": 0.8872 }
]
01
Vector-Native Indexing
Perform high-dimensional approximate nearest neighbor searches directly inside standard relational SQL tables.
02
ACID-Safe Vector Writes
Unlike stand-alone vector indices, Nexus guarantees complete transactional safety across both key-value records and vector node weights.
03
Infinite Horizontal Scale
Horizontal shard balancing happens auto-dynamically at the query router layer with zero manual shard layout maps.