AI Cluster Assistant

The AI Cluster Assistant is an AI-powered helper embedded in the Atlas UI. You can use it while creating a cluster or editing a cluster's configuration to answers natural-language questions about cluster configuration. The AI Cluster Assistant can then recommend settings based on your workload and needs and apply those settings directly to your cluster.

Enable AI Features

To use the AI Cluster Assistant in Atlas, first ensure that you enable generative AI features in the Organization Settings page.

To enable this organization-level setting:

In Atlas, go to the Organization Settings page.

If it's not already displayed, select your desired organization from the Organizations menu in the navigation bar.
In the sidebar, click Organization Settings.

The Organization Settings page displays.

Enable generative AI features.

Set the Enable Atlas features that use generative AI toggle to On.

To learn more, see Generative AI FAQs.

If generative AI features are enabled at the organization level, the AI Cluster Assistant is enabled for all projects by default. You can change this per project using the Atlas Administration API. To do so, update your project settings to toggle isClusterAiAssistantEnabled to false. If generative AI features are disabled at the organization level, the AI Cluster Assistant will be unavailable for all projects, even if isClusterAiAssistantEnabled is set to true for a project.

Access the AI Cluster Assistant

You can access the AI Cluster Assistant in the Atlas UI while creating a cluster or editing a cluster's configuration. To access the AI Cluster Assistant while creating a cluster:

In Atlas, go to the Clusters page for your project.

If it's not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
In the sidebar, click Clusters under the Database heading.

The Clusters page displays.

Open the Deploy your cluster page.

If you already have one or more clusters, click Create to display the Deploy your cluster page.

If this is your first cluster, click Build a Cluster to display the Deploy your cluster page.

Navigate to the AI Cluster Assistant.

Navigate to the bottom of the page and click Go to Advanced Configuration.

The AI Cluster Assistant chat window appears on the right side of the page.

To access AI Cluster Assistant while editing a cluster, simply click Edit additional configuration while on the Upgrade <yourClusterName> page.

To close the assistant, click the in the top right of the chat window. To reopen the assistant, click the . Closing the panel doesn't change your configuration or delete your chat history from that session. You can reopen the chat window at any time during the same session.

Example Flow

Navigate to the AI Cluster Assistant.

Follow the steps in Access the AI Cluster Assistant.

Ask a question about cluster configurations.

Submit a prompt looking for advice on your cluster configurations. You can discuss your workload, use case, or current pain points to have the AI Cluster Assistant suggest new configurations to best accommodate your needs.

Review the AI Cluster Assistant's response.

The AI Cluster Assistant provides a response discussing your needs and suggesting cluster configurations. Review the response and suggested configurations and evaluate if they are satisfactory.

Ask a follow-up question.

If needed, you can submit follow-up prompts to the AI Cluster Assistant to get more information on the suggested configurations, or to request new configurations that are better suited to your needs.

Click Apply these suggestions.

Click Apply these suggestions to accept the configurations suggested by the AI Cluster Assistant. This does not finalize your configurations and you can still make changes.

(Optional) Provide feedback.

You can use the or button under the AI Cluster Assistant response in the chat window to provide feedback on the response and your experience using the AI Cluster Assistant.

Save your cluster configurations.

Once satisfied, save your cluster configurations.

If you are creating a cluster, click Create Cluster at the bottom of the page to create your cluster with the suggested configurations.

If you are modifying a cluster, click Review Changes at the bottom of the page to see a summary of your changes. Then, click Accept Changes to save the new configurations.

This closes the AI Cluster Assistant.

While chatting with you, the AI Cluster Assistant offers links out to different Atlas UI pages, depending on your next task. The AI Cluster Assistant may link to one of these Atlas UI pages:

Examples

The following scenarios show some ways you can use the AI Cluster Assistant while creating or editing a cluster in Atlas.

Create a Cluster with the AI Cluster Assistant

Describe your use case and needs and ask the AI Cluster Assistant to suggest what cluster configurations could be best suited for you.

For example, if you are developing an application, you can tell the AI Cluster Assistant the purpose of your application and the expected amount of users.

I’m setting up a development cluster for a restaurant rating app with
100 users in NYC. What’s the best cluster configuration?

For a small development app with ~100 users, cost-effectiveness and
reliability are key. I recommend starting with an M10 cluster on AWS in
US_EAST_1 (N. Virginia). This tier supports your use case with enough
resources for development and light testing, while keeping costs
controlled. Avoid Free clusters (formerly known as M0) since you already have one and FLEX is
unnecessary for your scale.
I'll suggest a configuration based on these requirements. Quick
questions to further optimize:
- Do you expect your dataset to grow beyond 10GB during development?
- Any other environments needed, like staging or test?
- Will more developers or services connect to this cluster soon?
Suggesting the config now.

After the AI Cluster Assistant suggests a configuration, you can accept and save it or continue discussing your needs until you reach a configuration you are satisfied with.

You can also tell the AI Cluster Assistant the size of data you're expecting to need support for and the where your project is based.

I’m looking to migrate an existing project to Atlas. We currently
have 100GB of data that we need to move over. The project is currently
deployed in Europe and we want to maintain that. How should I configure
my cluster?

For migrating a 100GB project to Atlas in Europe, I recommend the following:
- Use AWS as your cloud provider (most commonly used, but let me know if
  you prefer GCP or Azure).
- Select the region EU_WEST_1 (Ireland) to host your data in Europe.
- Choose at least an M40 instance (this tier is recommended for
       migrating and running production workloads with 100GB+ data, ensuring
  strong performance during data load and normal operations).
- Set storage to 150GB (100GB data + 30% headroom).
- Default disk IOPS for M40 will provide ample performance for both
  migration and ongoing workload.
Let’s proceed with this setup.
A few quick clarifications:
- Is this a production environment with performance/uptime requirements?
- Do you expect rapid growth in data or traffic soon?
- Any preference for AWS, GCP, or Azure within Europe, or should we
  proceed with AWS?
I’ll configure the cluster now with these optimized settings.

Again, you can accept and save the suggested configuration or continue working towards a configuration that better meets your needs with the AI Cluster Assistant.

Modify a Cluster with the AI Cluster Assistant

Describe changes in your use case and needs and have the AI Cluster Assistant suggest how to best modify your cluster configurations.

For example, you can discuss budget constraints with the AI Cluster Assistant.

My cluster costs more than expected. Is there a setup that keeps
us under $500/mo?

To keep monthly costs under $500, I recommend switching to an M30 tier.
This provides reliable performance for standard production workloads at
around $394/month on AWS, plus some buffer for backups and storage.
Here’s how I’ll configure it:
- Instance size: M30 (entry-level dedicated, good for most production
  apps)
- Region: N. Virginia (US_EAST_1) (stays consistent with your existing
  clusters)
- Disk size: 40GB (default for M30, suitable for moderate workloads)
- No additional IOPS or premium storage (keeps costs down)
- No analytics/search node or advanced add-ons
Let’s proceed with this cost-effective configuration.
Before finalizing, can you clarify:
- Is this cluster for production, staging, or development?
- What’s your estimated data volume (just rough amount of GB needed)?
- Any specific performance requirements (high writes, reads, etc.)?
I’ll now suggest the configuration for you.

The AI Cluster Assistant suggests a new configuration.

You can also discuss increased workloads with the AI Cluster Assistant.

Suggest a cluster setup as I'm expanding to India and expect 20,000
new users and the associated data.

Since you're expanding to India with 20,000 expected users, you’ll need
a production-grade cluster deployed in an appropriate Indian region for
low latency. I recommend AWS as the cloud provider (unless you have a
preference for GCP or Azure), using the AP_SOUTH_1 region (Mumbai).
For your user base and workload size, an M30 dedicated tier is a strong
fit—it’s ideal for mid-sized production apps. This setup will handle
moderate to high traffic and provides headroom as your user base grows.
I'll set:
- Cloud Provider: AWS
- Region: AP_SOUTH_1 (Mumbai, India)
- Instance Size: M30 (production-ready with sufficient resources)
- Disk Size: Default (40GB, upgrade later if data grows quickly)
To refine this setup:
- What type of workload will the cluster handle (e.g., read-heavy,
  write-heavy, analytics)?
- Any data volume estimate (GB/TB) for the first year?
- Any specific cost constraints?
Applying this recommendation now.

The AI Cluster Assistant suggests a configuration that can accommodate your scaled needs.

Learn More

Data Usage for AI Cluster Assistant

Back

Create a Cluster

Create a Global Cluster