Eight practical patterns for explainable retrieval where agents already run — from browsers and robots to governed enterprise infrastructure.
01
Complete memory system for autonomous agents: semantic (vectors), episodic (graph), and structured data in one unified store.
memory.remember("User prefers concise answers")
context = memory.recall("response preferences")
print(memory.why("why concise answers?"))02
Combine knowledge graph traversal with vector similarity for superior context retrieval. MATCH + NEAR in one query.
SELECT fact.*, similarity()
FROM memories fact
WHERE vector NEAR $query
AND MATCH (fact)-[:SUPPORTED_BY]->(source);03
Build offline-capable AI applications with Tauri or Electron. Single binary, no server needed.
const results = await invoke("plugin:velesdb|search", {
request: { collection: "memories", vector: embedding, topK: 10 }
});04
Native SDKs for mobile with 32x memory compression via Binary Quantization.
let db = try VelesDatabase.open(path: "agent.db")
let collection = try db.getCollection(name: "memories")!
let results = try collection.search(vector: embedding, limit: 10)05
Run vector search directly in the browser with WASM. Privacy-first, no backend required.
import init, { VectorStore } from "@wiscale/velesdb-wasm";
await init();
const store = new VectorStore(384, "cosine");06
Keep vector and graph retrieval local to autonomous systems, without a mandatory network round trip.
SELECT * FROM observations
WHERE vector NEAR $sensor_embedding
LIMIT 5;07
Keep memory within controlled infrastructure and build the governance controls required by your regulatory program.
velesdb-server --data-dir /secure/memory --host 127.0.0.108
Share tenant-scoped memory between agents and reconstruct the state available at a past decision time with Premium forensic replay.
await fetch("/api/v1/memory/replay", {
method: "POST",
body: JSON.stringify({ decision, at, scope: "shared" })
}); // Premium