Mem0 Installation Guide

Ubuntu 24.04 LTS · Docker

What is Mem0?

Mem0 is a memory layer for AI agents — remembers user preferences, facts, and context across sessions. Gives LLM apps persistent memory without fine-tuning or vector-DB plumbing.

Prerequisites

• 2 CPU (4+ prod) · 2 GB RAM (4 GB+ local embed) · 5 GB+ disk
• Docker ≥ 24.0 + Compose · Port 8083 free · OpenAI key or local LLM

1. Docker Install

curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER

2. docker-compose.yml

mkdir -p ~/mem0 && cd ~/mem0
cat > docker-compose.yml << 'EOF'
services:
  mem0:
    image: mem0ai/mem0:latest
    container_name: mem0
    restart: unless-stopped
    ports:
      - "8083:8000"
    volumes:
      - ./mem0_data:/app/data
    environment:
      OPENAI_API_KEY: "${OPENAI_API_KEY:-}"
      MEM0_EMBEDDING_PROVIDER: "${MEM0_EMBEDDING_PROVIDER:-openai}"
      MEM0_CONFIG_PATH: "/app/data/mem0_config.yaml"
      MEM0_LOG_LEVEL: "${MEM0_LOG_LEVEL:-info}"
EOF
echo 'OPENAI_API_KEY=*** > .env

3. Launch

docker compose up -d
docker compose logs -f
curl http://localhost:8083/health

4. Nginx Proxy

sudo apt install -y nginx certbot python3-certbot-nginx

Create site proxying 127.0.0.1:8083 with standard headers and proxy_read_timeout 90s:

ln -s /etc/nginx/sites-available/mem0 /etc/nginx/sites-enabled/
nginx -t && systemctl reload nginx
certbot --nginx -d mem0.example.com

5. Env Vars

Var	Req	Def	Description
OPENAI_API_KEY	Yes*	—	OpenAI LLM+embed key
MEM0_EMBEDDING_PROVIDER	No	openai	openai,ollama,sentence-transformers,huggingface
MEM0_LLM_PROVIDER	No	openai	openai,ollama,anthropic,groq,litellm
MEM0_GRAPH_STORE	No	None	KG:neo4j,memgraph
MEM0_VECTOR_STORE	No	Embedded	Vector DB:chromadb,qdrant,pinecone,weaviate
MEM0_CONFIG_PATH	No	/app/data/conf.yaml	YAML config path
MEM0_LOG_LEVEL	No	info	debug,info,warning,error

* Omit with local LLM + local embeddings.

6. Local Embeddings

sentence-transformers

MEM0_EMBEDDING_PROVIDER: "sentence-transformers"
MEM0_EMBEDDING_MODEL: "all-MiniLM-L6-v2"

Ollama

MEM0_EMBEDDING_PROVIDER: "ollama"
MEM0_EMBEDDING_MODEL: "nomic-embed-text"
MEM0_OLLAMA_BASE_URL: "http://host.docker.internal:11434"

Add extra_hosts: ["host.docker.internal:host-gateway"]. ollama pull nomic-embed-text.

7. Backup

cd ~/mem0 && docker compose stop
tar -czf "mem0_$(date +%Y%m%d).tar.gz" mem0_data/
docker compose start

Schedule daily cron backing up ./mem0_data off-machine.

8. Update

cd ~/mem0 && docker compose pull && docker compose up -d --force-recreate
docker image prune -f

9. Troubleshooting

API Key

Symptom: 401. Fix: Check .env, no spaces, enough quota.

Embedding Errors

Symptom: EmbeddingError. Fix: Internet for first download. Ollama running. Increase Nginx timeout.

Retrieval Failing

Symptom: Empty results. Fix: Match embed provider. MEM0_LOG_LEVEL=debug.

Vector DB

Symptom: Connection refused. Fix: Check credentials. docker exec mem0 ping <host>. Data writable.

---

Mem0 on GitHub · Ubuntu 24.04