mem0-express-api

A TypeScript Express REST API that demonstrates how to add persistent memory to an AI assistant using Mem0 and Groq.

What it does

Most AI APIs are stateless — every request starts from zero. This project shows how to fix that by integrating Mem0 as a memory layer, so your AI assistant remembers users across conversations.

Stack

• Express.js — REST API framework
• TypeScript — type safety
• Mem0 — persistent memory layer
• Groq — LLM inference (llama-3.3-70b-versatile)

Endpoints

Method	Endpoint	Description
POST	/chat/:userId	Send a message and get a memory-aware response
GET	/chat/:userId/memories	Retrieve all stored memories for a user
DELETE	/chat/:userId/memories	Clear all memories for a user
POST	/chat/:userId/no-memory	Send a message with no memory context (for comparison)

Getting started

Prerequisites

• Node.js 18+
• Mem0 API key — mem0.ai/platform
• Groq API key — console.groq.com

Installation

git clone https://github.com/yourusername/mem0-express-api.git
cd mem0-express-api
npm install

Environment variables

Create a .env file in the root:

GROQ_API_KEY=your_groq_key_here
MEM0_API_KEY=your_mem0_key_here
PORT=3000

Run

npm run dev

Usage

Send a message:

curl -X POST http://localhost:3000/chat/user123 \
  -H "Content-Type: application/json" \
  -d '{"message": "My name is Ayomide and I am building an API platform called APIblok"}'

Check stored memories:

curl http://localhost:3000/chat/user123/memories

Compare with no memory:

curl -X POST http://localhost:3000/chat/user123/no-memory \
  -H "Content-Type: application/json" \
  -d '{"message": "What do you know about me?"}'

Clear memories:

curl -X DELETE http://localhost:3000/chat/user123/memories

How it works

1. User sends a message to /chat/:userId
2. Mem0 searches for relevant memories for that user
3. Memories are injected into the system prompt
4. Groq generates a context-aware response
5. The exchange is stored in Mem0 for future requests

Article

Read the full tutorial: [link to article]

License

MIT