Skip to main content

​
Agent Integration Examples

​
Example 1: Add Memory from Chat Messages (Memory Acquisition)

Store every user message as a memory for future context retrieval: 
// npm install @memorymodel/client
import { MemoryClient } from '@memorymodel/client';
import { generateText } from 'ai'; // Vercel AI SDK example

const memory = new MemoryClient({
  apiKey: process.env.MEMORY_API_KEY!,
  defaultEndUserId: "user_123"
});

async function handleUserMessage(message: string) {
  // 1. Store the message as memory
  await memory.add(message, {
    userContext: "User chat message"
  });

  // 2. Generate AI response (your LLM logic)
  const response = await generateText({
    model: openai('gpt-4o'),
    prompt: message
  });

  return response.text;
}

​
Example 2: Context-Aware Agent (Search Memory)

Search relevant memories before generating a response: 
// npm install @memorymodel/client
import { MemoryClient } from '@memorymodel/client';
import { generateText } from 'ai'; // Vercel AI SDK example

const memory = new MemoryClient({
  apiKey: process.env.MEMORY_API_KEY!,
  defaultEndUserId: "user_123"
});

async function handleUserQuery(query: string) {
  // 1. Search for relevant memories
  const memories = await memory.search(query, {
    limit: 5
  });

  // 2. Build context from memories
  const context = memories
    .map(m => `- ${m.content}`)
    .join('\n');

  // 3. Generate response with context
  const response = await generateText({
    model: openai('gpt-4o'),
    system: `You are a helpful assistant. Use this context:\n${context}`,
    prompt: query
  });

  return response.text;
}

​
Example 3: Full Agent Loop (Add Memory + Search Memory)

Complete pattern for a memory-powered chatbot: 
// npm install @memorymodel/client
import { MemoryClient } from '@memorymodel/client';

const memory = new MemoryClient({
  apiKey: process.env.MEMORY_API_KEY!,
  defaultEndUserId: "user_123"
});

async function agentLoop(userMessage: string) {
  // A. Add Memory (the new message)
  await memory.add(userMessage);

  // B. Search Memory (relevant past context)
  const context = await memory.search(userMessage, {
    limit: 3
  });

  // C. Generate response with memory context
  const systemPrompt = context.length > 0
    ? `Context from past conversations:\n${context.map(m => m.content).join('\n')}`
    : "No prior context available.";

  const aiResponse = await yourLLM.generate({
    system: systemPrompt,
    user: userMessage
  });

  // D. Optionally Add Memory (the AI response) too
  await memory.add(aiResponse, {
    userContext: "AI response"
  });

  return aiResponse;
}