Features Overview

Local-First Architecture

All data stays on your machine. No external calls, no cloud dependencies, works completely offline.

Built-in Embeddings

Use Embed() directly in queries — no external embedding service needed.

RAG Tooling

Vector search, BM25 keyword search, graph traversals, hybrid search with RRF.

Security

Private by default. Data accessible only through compiled SensibleQL queries.

Ultra-Low Latency

Rust + LMDB memory-mapped B-trees for near-zero overhead access.

Type-Safe Queries

Compile-time validation catches errors before production.

Multi-Model

Graph, vector, KV, document, and relational — all in one database.

MCP Server (Coming Soon)

Connect AI agents directly via MCP (Model Context Protocol) for agent memory and knowledge base use cases.