Use existing MongoDB instead of Postgres with LangGraph Standalone Container

[mhasan502]

We are currently using this docker-compose.yml to deploy AI agents. This is what the file looks like

services:
  langgraph-redis:
    image: redis:6
    restart: always
    ports:
      - "6379:6379"
    healthcheck:
      test: redis-cli ping
      interval: 5s
      timeout: 1s
      retries: 5

  langgraph-postgres:
    image: pgvector/pgvector:pg16
    restart: always
    ports:
      - "5433:5432"
    environment:
      POSTGRES_DB: postgres
      POSTGRES_USER: ${POSTGRES_USER:-postgres}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-postgres}
    command:
      - postgres
      - -c
      - shared_preload_libraries=vector
    volumes:
      - langgraph-data:/var/lib/postgresql/data
    healthcheck:
      test: pg_isready -U postgres
      start_period: 10s
      timeout: 1s
      retries: 5
      interval: 60s
      start_interval: 1s

  langgraph-api:
    build:
      context: .
      dockerfile: Dockerfile
    restart: always
    ports:
      - "8123:8000"
    depends_on:
      langgraph-redis:
        condition: service_healthy
      langgraph-postgres:
        condition: service_healthy
    env_file:
      - .env
    healthcheck:
      test: python /api/healthcheck.py
      interval: 60s
      start_interval: 1s
      start_period: 10s


volumes:
  langgraph-data:
    driver: local

Our application, excluding AI agents, uses MongoDB Atlas. We want to move away from Postgres. Since LangGraph already supports MongoDB, how can I set up MongoDB instead of Postgres?

[nikhildigde]

In the same situation. Do we need to package the langgraph server in our own image (instead of the langraph build generated one that add a lot of deps etc.)? And then use the mongodbsaver in the agent?