At a glance
| Database engines | MongoDB (replica sets and sharded clusters) |
| Regions | More than 125 regions across AWS, Azure and Google Cloud |
| Scaling | Change tier, auto-scaling within bounds, sharding on Dedicated clusters; Atlas Infinite (preview) scales compute and storage separately |
| Backups and recovery | Free: no Atlas backups; Flex: daily snapshots, last 8 kept; M10+: Cloud Backups with point-in-time restore |
| High availability | Replica sets of at least three nodes; electable nodes across zones, regions or clouds |
| Free tier | Yes: Free cluster with 512 MB, shared compute, one per project |
| Pricing model | Free; Flex hourly by ops/sec, capped at 30 USD a month; Dedicated hourly per tier plus backups, transfer and extras |
MongoDB Atlas is the default way to run MongoDB in the cloud if you want the full MongoDB feature set without operating replica sets yourself. It starts free, and its multi-region and multi-cloud clusters are hard to match elsewhere; costs grow with tier, backups, data transfer and extra nodes, so price your exact provider and region.
Overview
Atlas is the managed cloud service from MongoDB, Inc., the company that develops MongoDB. You pick a cloud provider, region and tier, and Atlas deploys, patches, monitors and backs up the cluster. Every Atlas cluster is a replica set of at least three nodes, and Dedicated clusters can be sharded. MongoDB says Atlas runs in more than 125 regions across AWS, Azure and Google Cloud, and one cluster can span several regions and providers.
MongoDB now groups Atlas into Atlas Core (Free, Flex and Dedicated tiers for predictable workloads) and Atlas Infinite, a public-preview option for M10 and larger clusters that scales compute and storage separately and does not yet support multi-region clusters. Alongside the database, Atlas includes full-text and vector search and stream processing. Ways to host MongoDB outside Atlas are covered in our MongoDB hosting guide.
Key features
- Replica sets of at least three nodes, with 3, 5 or 7 electable nodes that can be spread across regions and cloud providers
- Multi-region and multi-cloud clusters on AWS, Azure and Google Cloud (Atlas Core)
- Cloud Backups on M10+ clusters using the provider's snapshots, with restore to a chosen date and time or oplog entry from Continuous Cloud Backup
- Read-only and analytics nodes (analytics nodes on M10+) to isolate reporting queries from the operational workload
- Auto-scaling of cluster resources within bounds you set
- Built-in full-text search, vector search and stream processing
- Encryption at rest, in transit and in use, and management through the UI, the Atlas CLI or APIs
Pros and cons
Pros
- Run by MongoDB, Inc., the company that develops the MongoDB server, so it is MongoDB itself rather than a compatible engine
- One cluster can span regions and AWS, Azure and Google Cloud
- Free and capped-price Flex tiers make it cheap to start
- Search, vector search and stream processing without a separate service
Cons
- Free clusters have no Atlas backups and are limited to 512 MB, one per project
- Flex backups are daily snapshots only; point-in-time restore needs an M10 or larger cluster
- Dedicated costs grow with backups, data transfer, extra nodes and regions, and differ by provider and region
- Atlas Infinite is in public preview and does not support multi-region or multi-cloud clusters yet
Pricing
Atlas Core has three tiers: Free (0 USD, 512 MB), Flex (from 0.011 USD per hour, capped at 30 USD a month) and Dedicated clusters from M10 (0.08 USD per hour, which MongoDB shows as starting at 56.94 USD a month), in USD on the AWS tab of the pricing page. Dedicated clusters are billed hourly and invoiced monthly, with backups, data transfer, extra storage, nodes and regions on top; prices vary by cloud provider and region. Atlas Infinite, in public preview, starts at 0.09 USD per hour. For tier tables and worked monthly examples, see our MongoDB Atlas pricing guide.
Pricing checked on the vendor's official page on 7 November 2026. Prices change; confirm before buying.
Who it is best for
Teams building on MongoDB that want a managed service with the full feature set, from a free cluster for prototypes to multi-region production clusters.
Who should look elsewhere
- Relational workloads that need joins and SQL: see MongoDB vs PostgreSQL.
- Teams committed to staying inside one cloud provider's native services: see DynamoDB vs MongoDB and Azure Cosmos DB.
Verdict
MongoDB Atlas is the most complete way to run MongoDB as a service, with multi-cloud clusters, built-in search and a free starting tier. Plan for production on M10 or larger if you need point-in-time restore, and use the pricing calculator for your provider and region, because the extras often matter as much as the tier price.
Frequently asked questions
Is MongoDB Atlas free?
Atlas has a Free tier that costs nothing indefinitely, with 512 MB of storage and shared compute, one cluster per project and no Atlas backups. Flex starts at 0.011 USD per hour and is capped at 30 USD a month. Our Atlas pricing guide has the full tier list.
What replaced Atlas Serverless and the M2 and M5 tiers?
Flex clusters. MongoDB stopped new M2, M5 and Serverless deployments in February 2025, and from 22 January 2026 Atlas no longer supports them; existing instances were migrated.
Can an Atlas cluster run on more than one cloud?
Yes, on Atlas Core. You can place nodes in several regions of one provider or across AWS, Azure and Google Cloud, as long as the total number of electable nodes is 3, 5 or 7. Atlas Infinite does not support multi-region or multi-cloud clusters during its public preview.
Sources
- MongoDB Atlas database
- MongoDB pricing
- Atlas docs: Cloud Backups
- Atlas docs: Flex cluster backups
- Atlas docs: High availability and workload isolation
- Atlas docs: Flex cluster costs
Checked November 2026.
How we research reviews and what star ratings mean: our editorial method.