mem0 | Dosu

Overview#

A personalized AI chat experience powered by Mem0 that remembers user preferences, facts, and context across conversations. Unlike traditional chat applications where each conversation starts fresh, this example maintains a persistent memory of user interactions to provide increasingly relevant and personalized responses.

Features#

Persistent Memory: Facts and preferences survive across chat sessions
Automatic Extraction: AI automatically identifies and stores important information
Context Injection: Relevant memories are included in each conversation
Memory Management: Users can view, edit, and delete stored memories
Privacy Controls: Fine-grained control over what gets remembered
Memory Visualization: Clear interface showing what the AI remembers

How It Works#

Conversation: User chats normally with the AI
Extraction: Mem0 identifies key facts (e.g., "User prefers dark mode")
Storage: Facts are stored in Mem0's memory system
Retrieval: On future conversations, relevant memories are retrieved
Personalization: AI responses incorporate remembered context

Integration#

The example integrates Mem0's memory SDK with assistant-ui:

import MemoryClient from "mem0ai";

const mem0 = new MemoryClient({
  apiKey: process.env.MEM0_API_KEY,
});

// Store memories from a conversation
async function storeMemory(userId: string, messages: Message[]) {
  await mem0.add(
    messages.map((m) => ({
      role: m.role,
      content: m.content,
    })),
    { user_id: userId },
  );
}

// Retrieve relevant memories for context
async function getRelevantMemories(userId: string, query: string) {
  const { results } = await mem0.search(query, {
    user_id: userId,
    limit: 10,
  });

  return results.map((m) => m.memory).join("\n");
}

Using Memories in Chat#

// In your API route
export async function POST(req: Request) {
  const { messages, userId } = await req.json();
  const lastMessage = messages[messages.length - 1].content;

  // Get relevant memories
  const memories = await getRelevantMemories(userId, lastMessage);

  // Include memories in system prompt
  const systemPrompt = `You are a helpful assistant.

Here's what you remember about this user:
${memories}

Use this context to personalize your responses.`;

  const result = streamText({
    model: openai("gpt-5.6-luna"),
    system: systemPrompt,
    messages,
  });

  // Store new memories after response
  await storeMemory(userId, messages);

  return result.toUIMessageStreamResponse();
}

Memory Types#

Type	Example
Preferences	"User prefers concise answers"
Facts	"User works at Acme Corp"
Context	"User is building a React app"
History	"Previously discussed authentication"

Source#

Overview#

Features#

Persistent Memory: Facts and preferences survive across chat sessions

Automatic Extraction: AI automatically identifies and stores important information

Context Injection: Relevant memories are included in each conversation

Memory Management: Users can view, edit, and delete stored memories

Privacy Controls: Fine-grained control over what gets remembered

Memory Visualization: Clear interface showing what the AI remembers

Integration#

The example integrates Mem0's memory SDK with assistant-ui:

import MemoryClient from "mem0ai";

const mem0 = new MemoryClient({
  apiKey: process.env.MEM0_API_KEY,
});

// Store memories from a conversation
async function storeMemory(userId: string, messages: Message[]) {
  await mem0.add(
    messages.map((m) => ({
      role: m.role,
      content: m.content,
    })),
    { user_id: userId },
  );
}

// Retrieve relevant memories for context
async function getRelevantMemories(userId: string, query: string) {
  const { results } = await mem0.search(query, {
    user_id: userId,
    limit: 10,
  });

  return results.map((m) => m.memory).join("\n");
}

Using Memories in Chat#

// In your API route
export async function POST(req: Request) {
  const { messages, userId } = await req.json();
  const lastMessage = messages[messages.length - 1].content;

  // Get relevant memories
  const memories = await getRelevantMemories(userId, lastMessage);

  // Include memories in system prompt
  const systemPrompt = `You are a helpful assistant.

Here's what you remember about this user:
${memories}

Use this context to personalize your responses.`;

  const result = streamText({
    model: openai("gpt-5.6-luna"),
    system: systemPrompt,
    messages,
  });

  // Store new memories after response
  await storeMemory(userId, messages);

  return result.toUIMessageStreamResponse();
}

Memory Types#

Type	Example
Preferences	"User prefers concise answers"
Facts	"User works at Acme Corp"
Context	"User is building a React app"
History	"Previously discussed authentication"