MemOS Helm Chart

MemOS - AI Memory Operating System for LLM and Agent systems.

Prerequisites

• Kubernetes 1.20+
• Helm 3.0+
• PV provisioner support in the underlying infrastructure

Components

Component	Description	Port
memos-api	Main API service	8000
neo4j	Graph database	7474 (HTTP), 7687 (Bolt)
qdrant	Vector database	6333 (HTTP), 6334 (gRPC)

Quick Start

1. Build Docker Image

# From repo root
docker build -t memos:2.0.9 .

2. Push to Registry

docker tag memos:2.0.9 your-registry/memos:2.0.9
docker push your-registry/memos:2.0.9

3. Configure Values

cp deploy/helm/values-example.yaml my-values.yaml

# Edit my-values.yaml with your settings:
# - OPENAI_API_KEY
# - MEMRADER_API_KEY
# - image.repository (your registry)

4. Install

helm install memos deploy/helm -f my-values.yaml -n memos --create-namespace

Configuration

Required Settings

memos:
  image:
    repository: your-registry/memos
    tag: "2.0.9"
  env:
    OPENAI_API_KEY: "sk-your-key"
    MEMRADER_API_KEY: "sk-your-key"

Enable Ingress

ingress:
  enabled: true
  className: nginx
  hosts:
    - host: memos.yourdomain.com
      paths:
        - path: /
          pathType: Prefix

Use External Neo4j/Qdrant

neo4j:
  enabled: false

qdrant:
  enabled: false

memos:
  env:
    NEO4J_URI: "bolt://external-neo4j:7687"
    NEO4J_USER: "neo4j"
    NEO4J_PASSWORD: "password"
    QDRANT_HOST: "external-qdrant"
    QDRANT_PORT: "6333"

Values Reference

Key	Default	Description
memos.replicaCount	1	API replicas
memos.image.repository	memos/memos	Image repository
memos.image.tag	2.0.9	Image tag
memos.service.type	ClusterIP	Service type
memos.service.port	8000	Service port
neo4j.enabled	true	Enable Neo4j
neo4j.auth.password	memos123456	Neo4j password
neo4j.persistence.size	10Gi	Neo4j data size
qdrant.enabled	true	Enable Qdrant
qdrant.persistence.size	10Gi	Qdrant data size
ingress.enabled	false	Enable Ingress

API Endpoints

# Add memory
curl -X POST http://memos-api:8000/product/add \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": "test-user",
    "mem_cube_id": "test-cube",
    "messages": [{"role": "user", "content": "I like strawberry"}],
    "async_mode": "sync"
  }'

# Search memory
curl -X POST http://memos-api:8000/product/search \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What do I like",
    "user_id": "test-user",
    "mem_cube_id": "test-cube"
  }'

Uninstall

helm uninstall memos -n memos
kubectl delete namespace memos

Troubleshooting

# Check logs
kubectl logs -n memos -l app.kubernetes.io/component=api

# Check Neo4j
kubectl exec -n memos -it deployment/memos-neo4j -- cypher-shell -u neo4j -p memos123456

# Check Qdrant
kubectl exec -n memos -it deployment/memos-qdrant -- curl localhost:6333