On-Prem Hosting with MongoDB vector search

MongoDB Atlas Vector Search
1

Hi MongoDB Community,

I am facing an issue where my MongoDB database has to be hosted on-prem within a private cloud service, hence I am unable to use Atlas for hosting my database as it would be outside the private cloud cluster. My main bottleneck is in finding a way to enable native vector search for RAG without Atlas, because from my understanding vector search is only enabled for Atlas-hosted databases in production. Any way to host Atlas on-prem within a private cluster or enable vector search functionality natively within MongoDB (without using Atlas / Atlas CLI) in prod?

3

Hi @Yi_Khuen_Chai,

Thanks for joining the MongoDB community, welcome here! Full text search & Vector search are coming to MongoDB Community Edition. You can register for a preview here:

But I’ve been already running containers using MongoDB Local Atlas clusters, the same ones that you can install using atlas-cli. To see that in action, create a new Codespace in this repo: GitHub - mongodb-developer/jedee: Jupyter EDEE. This will spin two containers, one running a MongoDB instance that supports Full Text Search and Vector Search and a 2nd one to host a bunch of Jupyter notebooks (to run code).

Long story short, this Docker Compose file has the configuration for both containers: jedee/.devcontainer/docker-compose.yml at main · mongodb-developer/jedee · GitHub

Relevant snippet:

  mongodb:
    image: mongodb/mongodb-atlas-local:8.0.3-20250506T093411Z
    restart: unless-stopped
    volumes:
      - mongodb-data:/data/db

    # Uncomment to change startup options
    environment:
      - MONGODB_INITDB_ROOT_USERNAME=admin
      - MONGODB_INITDB_ROOT_PASSWORD=mongodb
      - MONGO_INITDB_DATABASE=library
    ports:
      - 27017:27017

I think that’s what you need. You can even run Docker inside your cloud-hosted machine. It works :wink: