Adding Memory

By default, agents are stateless. Adding memory lets your agent remember previous messages in a session.

KV Memory (session history)

Best for: conversation history that doesn’t need to persist beyond a few days.

import { Agent } from '@moon-wave/core';
import { KVMemoryAdapter } from '@moon-wave/memory';

interface Env {
  GROQ_API_KEY: string;
  SESSIONS: KVNamespace;
}

export default {
  async fetch(request: Request, env: Env): Promise<Response> {
    const agent = new Agent({
      name: 'my-agent',
      model: { provider: 'groq', model: 'llama-3.3-70b-versatile' },
      systemPrompt: 'You are a helpful assistant.',
      memory: {
        type: 'kv',
        adapter: new KVMemoryAdapter(env.SESSIONS),
      },
    });

    const { sessionId, input } = await request.json() as { sessionId: string; input: string };
    const result = await agent.run(input, { sessionId, env });
    return Response.json({ output: result.output });
  },
};

Add the KV binding to wrangler.toml:

[[kv_namespaces]]
binding = "SESSIONS"
id = "your-kv-namespace-id"

Create it:

npx wrangler kv namespace create SESSIONS

D1 Memory (persistent)

Best for: long-term conversation history stored in SQL.

import { D1MemoryAdapter } from '@moon-wave/memory';

const agent = new Agent({
  // ...
  memory: {
    type: 'd1',
    adapter: new D1MemoryAdapter(env.DB),
  },
});

Run the migration:

npx wrangler d1 execute my-db --file=./node_modules/@moon-wave/memory/migrations/001_init.sql

Add to wrangler.toml:

[[d1_databases]]
binding = "DB"
database_name = "my-agent-db"
database_id = "your-d1-id"

Vectorize Memory (semantic search)

Best for: agents that need to recall relevant past context, not just recent messages.

import { VectorizeAdapter, MemoryManager } from '@moon-wave/memory';

const memory = new MemoryManager({
  shortTerm: new KVMemoryAdapter(env.SESSIONS),
  longTerm: new D1MemoryAdapter(env.DB),
  vector: new VectorizeAdapter(env.VECTORIZE, env.AI),
});

// Store a fact
await memory.remember('User prefers dark mode', { userId });

// Retrieve relevant facts
const facts = await memory.recall('user preferences', 3);