JavaScript SDK

Installation

npm install @memoid/sdk

Quick Start

import { MemoryClient } from "@memoid/sdk";

const client = new MemoryClient("your-api-key");

// Add a memory
await client.add([
  { role: "user", content: "I love hiking and Italian food" },
  { role: "assistant", content: "Great! I'll remember that." }
], { user_id: "user_123" });

// Search memories
const results = await client.search("What food does the user like?", {
  user_id: "user_123"
});

// Recall context (memories + graph in one call)
const context = await client.recall("Tell me about the user", {
  user_id: "user_123",
  include_graph: true
});

Configuration

const client = new MemoryClient("your-api-key", {
  baseUrl: "https://api.memoid.dev",  // default
  timeout: 30000                       // 30s default
});

For self-hosted instances, set baseUrl to your API URL.

Memories

Add Memories

const memories = await client.add(
  [
    { role: "user", content: "My manager Sarah works at TechCorp" }
  ],
  {
    user_id: "user_123",
    project_id: "proj_abc",
    metadata: { source: "chat" },
    extract_graph: true  // extract entities to knowledge graph
  }
);

Get a Memory

const memory = await client.get("memory_id");

List Memories

const memories = await client.getAll({
  user_id: "user_123",
  limit: 50
});

Update a Memory

await client.update("memory_id", "Updated memory text", { source: "edit" });

Delete

await client.delete("memory_id");
await client.deleteAll({ user_id: "user_123" });

History

const history = await client.history("memory_id");

Search

const results = await client.search("food preferences", {
  user_id: "user_123",
  limit: 5,
  threshold: 0.7
});

for (const result of results) {
  console.log(`${result.memory} (score: ${result.score})`);
}

Recall

Unified context retrieval — combines memory search with knowledge graph data in a single call. Ideal for agents.

const context = await client.recall("What do I know about John?", {
  user_id: "user_123",
  include_graph: true,
  memory_limit: 10,
  threshold: 0.3
});

console.log(context.memories);       // relevant memories
console.log(context.entities);       // related graph entities
console.log(context.relationships);  // entity relationships

User Context

Get all memories and graph data for a user in a single call — ideal for rendering a graph + memory list in the UI.

const context = await client.getUserContext({ user_id: "user_123" });

console.log(context.memories);       // all memories
console.log(context.entities);       // all graph entities
console.log(context.relationships);  // all relationships
console.log(context.stats);          // { total_memories, total_entities, total_relationships }

Knowledge Graph

Extract Knowledge

const result = await client.extractKnowledge(
  "Sarah is the CEO of TechCorp. She founded it with Mike.",
  { user_id: "user_123" }
);
// result.entities: [{name: "Sarah", type: "person"}, ...]
// result.relationships: [{subject: "Sarah", predicate: "ceo_of", object: "TechCorp"}, ...]

Add Entity & Relationship

await client.addEntity("Acme Corp", "organization");
await client.addRelationship("John", "person", "works_at", "Acme Corp", "organization");

Query & Search

const related = await client.graphQuery("John", 2);  // 2 hops
const found = await client.graphSearch("tech companies");

Delete Entity

await client.deleteEntity("John");  // also removes relationships

Batch Operations

await client.batchUpdate([
  { id: "mem_1", memory: "Updated text" },
  { id: "mem_2", memory: "Another update" }
]);

await client.batchDelete({ memory_ids: ["mem_1", "mem_2"] });

Projects

const project = await client.createProject("My Project");
const projects = await client.getProjects();
await client.updateProject("proj_id", { name: "Renamed" });
await client.deleteProject("proj_id");

Usage

const stats = await client.getUsageStats();
const daily = await client.getUsageDaily(30);

TypeScript

All methods are fully typed. Import types directly:

import type { Memory, SearchResult, RecallResult, GraphEntity } from "@memoid/sdk";