Skip to content
little big brain

What is little big brain?

little big brain (lbb) is an object-storage-native temporal graph and hybrid-search engine for AI context, written in Rust. It stores graph facts as append-only events, builds search indexes as disposable object-backed accelerators, and runs semantic hybrid search across BM25, vector, ontology, and graph-neighborhood signals — all over the same consistent snapshot.

It is designed to be the retrieval and memory layer behind AI applications and agents: durable enough to be the source of truth, and rich enough to answer lexical, semantic, and graph-shaped questions from one place.

What it is good at

Section titled “What it is good at”
  • Hybrid search that actually fuses signals. Lexical matching, BM25, vectors/ANN, ontology resolution, and graph-neighborhood traversal are combined and ranked together — with filters, facets, and reciprocal-rank fusion — not bolted on as separate systems.
  • A temporal knowledge graph. Facts are append-only observations with provenance and evidence. You can read current state, relationship history, and why a fact exists, and pin any query to a historical snapshot.
  • Standards-based query. A conformant SPARQL 1.1 engine, a structured SELECT/aggregate surface, and SHACL Core / SHACL-SPARQL / SHACL-AF validation and inference run over the same object-storage substrate.
  • Reasoning that derives what you didn’t write. RDFS type closure and SHACL-AF rule inference fold into ordinary queries — deterministic, transparent, and pinnable to a snapshot. See reasoning.
  • Object storage as the source of truth. The durable graph, WAL, and index runs live in object storage (local filesystem or S3-compatible). Compute is stateless and disposable; indexes are derived acceleration structures.

How you talk to it

Section titled “How you talk to it”
Surface Best for
TypeScript SDK (@lbb/client) Node, browser, and edge apps
Python SDK (lbb) Scripts, notebooks, data/ML workflows
MCP server (@lbb/mcp) Giving an AI agent a graph-aware tool belt
HTTP API Any language; also the native SPARQL 1.1 Protocol
Console Visual exploration, query, indexing, and operations
CLI Operating a self-hosted deployment (coming with self-hosting)

All of the application SDKs speak the same HTTP API and authenticate with the same stack API key, so you can mix them freely — write with Python, retrieve from TypeScript, let an agent read through MCP, and explore in the console, all against one graph.

Deployment shape

Section titled “Deployment shape”

Today little big brain is a hosted product: a managed data plane at db.eu.littlebigbrain.com, an account/stack API at api.littlebigbrain.com, a hosted MCP endpoint at mcp.littlebigbrain.com, and the console at cloud.littlebigbrain.com. You sign up, create a stack (an isolated tenant), and get a per-stack API key.

A self-hostable distribution — a Docker container / pre-built server that runs against a local filesystem or an S3-compatible bucket, with the CLI for operations — is on the roadmap.

Where to go next

Section titled “Where to go next”