Quickstart

This walks the core pipeline against the hosted data plane: write facts → build indexes → search. You’ll do it with the TypeScript or Python SDK, or a plain curl.

You need a stack and a stack API key. Create a stack in the console and copy its key — it looks like lbb_sk_live_… or lbb_sk_test_…. See Authentication & endpoints for the details.

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, // lbb_sk_live_… / lbb_sk_test_…
});

// 1. Write facts (triplets) into the default graph.
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,
        evidence: "auth-service writes identity records to user-db",
      },
    ],
  },
  { idempotencyKey: "quickstart-1" },
);

// 2. Build persisted BM25 + vector + adjacency indexes and wait for them.
await lbb.indexes.run({ wait: true });

// 3. Run hybrid search over the snapshot.
const results = await lbb.search.hybrid(
  "which systems store customer identity data",
  { topK: 5, source: "persisted", consistency: "strong", targets: ["entities", "assertions"] },
);
for (const a of results.assertions ?? []) {
  console.log(a.relation?.name, a.score);
}

pip install lbb-sdk

from lbb import LbbClient

with LbbClient("https://db.eu.littlebigbrain.com", api_key="lbb_sk_live_...") as lbb:
    # 1. Write facts.
    lbb.graph("main").facts.create({
        "triplets": [{
            "source": {"type": "SERVICE", "name": "auth-service"},
            "relation": "WRITES_TO",
            "target": {"type": "DATABASE", "name": "user-db"},
            "confidence": 0.93,
            "evidence": "auth-service writes identity records to user-db",
        }],
    }, idempotency_key="quickstart-1")

    # 2. Build indexes.
    lbb.indexes.run(wait=True)

    # 3. Hybrid search.
    results = lbb.search.hybrid(
        "which systems store customer identity data",
        top_k=5, source="persisted", consistency="strong",
        targets=["entities", "assertions"],
    )
    for a in results.get("assertions", []):
        print(a["relation"]["name"], a["score"])

export LBB_URL=https://db.eu.littlebigbrain.com
export LBB_KEY=lbb_sk_live_...

# 1. Write facts.
curl -sS -X POST "$LBB_URL/v1/graph/commit?graph=main" \
  -H "Authorization: Bearer $LBB_KEY" \
  -H "Idempotency-Key: quickstart-1" \
  -H "Content-Type: application/json" \
  -d '{"triplets":[{"source":{"type":"SERVICE","name":"auth-service"},
       "relation":"WRITES_TO","target":{"type":"DATABASE","name":"user-db"},
       "confidence":0.93,"evidence":"auth-service writes identity records to user-db"}]}'

# 2. Build indexes.
curl -sS -X POST "$LBB_URL/v1/index/run?graph=main&wait=true" -H "Authorization: Bearer $LBB_KEY"

# 3. Hybrid search.
curl -sS "$LBB_URL/v1/search?graph=main&source=persisted&query=which%20systems%20store%20customer%20identity%20data" \
  -H "Authorization: Bearer $LBB_KEY"

Prefer not to write code first? Do the same three steps visually in the console — load data, build indexes in Retrieval → Indexes, and run a search in Search & eval.

What just happened

Write appended immutable graph events (entities, a relation event, and the evidence that supports it) to the write-ahead log.
Index built persisted BM25, vector/ANN, and adjacency runs — disposable accelerators derived from the graph snapshot.
Search fused lexical + BM25 + vector + ontology + graph-neighborhood signals over one snapshot, with strong consistency overlaying any facts committed after the index run.

Read Core concepts to understand snapshots, consistency, and the append-only model, then head to the use-case guides.