overview
object-storage-native uptime —storage —runtime health
local process counters · object I/O · cache efficiency JSON + text exportRuntime health is not available in this console session. Open Little Big Brain Cloud to view activity for the selected stack.
—
—
— —ontology
concepts · terms · synonyms · relations · resolveingest
commit triplets into the WAL · idempotent per Idempotency-Key—
indexing
continuous · delta segments · merge-on-read indexing continuouslydatasets
built-in fixtures + downloaded public benchmark mirror.lbb-datasets/ is present.models
local graph-supervised embeddings · auto-research loopLBB_MODEL_TRAINING_ENABLED=true to enable the launcher.
Stored-provider query preview appears after a run is selected.
No logs yet.
storage
object-storage-native · read-only · immutableStorage is append-only and immutable — facts are never edited or deleted in place; indexes rebuild from the WAL. This view is read-only by design.
api & sdk
integrate · query · automateOne API — four ways in
Everything talks to one HTTP API. Authenticate with a stack API key as a bearer credential and scope each call to a graph / branch (default main). The OpenAPI spec is the single source of truth — the MCP server, both SDKs, and the CLI are generated from it, so they never drift from the server.
MCP — agent tool belt
Claude Code · Cursor · CodexGive an agent a graph-aware tool belt over your stack. Local (stdio) — drop this into your editor's MCP config (.mcp.json / mcp.json); npx fetches the server on demand:
{
"mcpServers": {
"lbb": {
"command": "npx",
"args": ["-y", "@lbb/mcp"],
"env": {
"LBB_BASE_URL": "https://db.eu.littlebigbrain.com",
"LBB_API_KEY": "lbb_sk_live_…"
}
}
}
}
Set LBB_GRAPH / LBB_BRANCH to target a non-default graph. The belt is task-shaped and compact by default; each tool dispatches to the right Little Big Brain API route:
lbb_searchnatural-language retrieval, fusion, and optional path followinglbb_inspectguide, ontology, metadata, state, history, why, traverselbb_querystructured query, SPARQL, SHACL, inference, analysislbb_commitwrite facts; omitted idempotency keys are content-derivedlbb_configuredefine ontology or stored inference ruleslbb_indexrefresh persisted indexes after a commit batchSDKs
generated from the OpenAPI spec# pip install lbb-sdk
from lbb import LbbClient
with LbbClient("https://db.eu.littlebigbrain.com", api_key="lbb_sk_live_…",
graph="main", branch="main") as lbb:
lbb.graph("main").facts.create({"triplets": [
{"source": {"type": "SERVICE", "name": "auth-service"},
"relation": "WRITES_TO",
"target": {"type": "DATABASE", "name": "user-db"},
"confidence": 0.93}]},
idempotency_key="import-1")
lbb.indexes.run(wait=True)
hits = lbb.search.hybrid("customer identity data", top_k=5)
// npm install @lbb/client
import { LbbClient } from "@lbb/client";
const lbb = new LbbClient({
baseUrl: "https://db.eu.littlebigbrain.com",
apiKey: process.env.LBB_API_KEY,
graph: "main", branch: "main",
});
await lbb.graph("main").facts.create({ triplets: [
{ source: { type: "SERVICE", name: "auth-service" },
relation: "WRITES_TO",
target: { type: "DATABASE", name: "user-db" },
confidence: 0.93 } ] }, { idempotencyKey: "import-1" });
await lbb.indexes.run({ wait: true });
const hits = await lbb.search.hybrid("customer identity data", { topK: 5 });
Sync + async (Python AsyncLbbClient), same method set across both: commit · search / graph_search / multi_search / full_text_search / embedding_search · traverse / semantic_traverse · current_state / history / why / shacl · ontology_* · index_* / compact · status / metadata.
HTTP API
curl · any language# commit a triplet curl -X POST 'https://db.eu.littlebigbrain.com/v1/graph/commit?graph=main&branch=main' \ -H 'Authorization: Bearer lbb_sk_live_…' -H 'Content-Type: application/json' \ -H 'Idempotency-Key: import-1' \ -d '{"triplets":[ {"source":{"type":"SERVICE","name":"auth-service"}, "relation":"WRITES_TO", "target":{"type":"DATABASE","name":"user-db"}, "confidence":0.93}]}' # build persisted indexes, then search curl -X POST 'https://db.eu.littlebigbrain.com/v1/index/run?graph=main' -H 'Authorization: Bearer lbb_sk_live_…' curl 'https://db.eu.littlebigbrain.com/v1/search?graph=main&query=customer%20identity%20data' \ -H 'Authorization: Bearer lbb_sk_live_…'
Common routes: POST /v1/graph/commit · GET /v1/search · POST /v1/graph/search · POST /v1/index/run · GET /v1/status · POST /v1/query/shacl · POST /v1/ontology/search. A write that violates a relation's declared domain/range comes back with a conformance warning (it never blocks). Full reference: the OpenAPI spec and docs/reference/server.md.
CLI
lbb-testctl · local & scriptingROOT=/tmp/lbb cargo run -p lbb-testctl -- graph create --root $ROOT --tenant acme --graph main --branch main cargo run -p lbb-testctl -- graph triplets commit --root $ROOT --tenant acme --graph main --branch main \ --file database/examples/triplets/semantic_graph.json cargo run -p lbb-testctl -- graph index --root $ROOT --tenant acme --graph main --branch main \ --file database/examples/search/search_index_run.json cargo run -p lbb-testctl -- graph search-eval --root $ROOT --tenant acme --graph main --branch main \ --file database/examples/perf/search_accuracy_smoke.json
Also: graph compact · graph metadata · storage conformance. Point it at S3 with --backend s3 --s3-bucket … --s3-endpoint …. Full surface: docs/reference/cli.md.
stacks
isolated tenants · per-stack api keys · control plane in object storageEach stack is an isolated tenant under its own object prefix; the API key alone selects the tenant. Rotation is immediate. Stack creation requires a verified email.
api keys
stack: —Keys are hashed and stored under control/v1/; only a hash is kept. Revocation is immediate — the key index record is read on every request. The backend models one active key per stack (rotated); multiple named per-stack keys are designed, not yet shipped.
usage & billing
free plan $0.00 dueBilling enforcement is not yet shipped — this view is informational; no quota gating or charge flow is active.