What is MongoDB Atlas?
MongoDB Atlas is a database-as-a-service that deploys and operates MongoDB clusters for its customers. Users choose a cloud provider, region and cluster tier in the web console, the Atlas CLI, the Administration API or infrastructure-as-code tools such as Terraform, and Atlas provisions the servers, configures replication, applies patches and version upgrades and can scale compute and storage automatically. Clusters are available in more than 125 regions on Amazon Web Services, Google Cloud and Microsoft Azure, and a single cluster can place its nodes with more than one of those providers.
Atlas also bundles services that would otherwise need separate systems. MongoDB Search maintains Apache Lucene indexes for full-text search, Vector Search indexes embeddings for semantic retrieval, Atlas Stream Processing works on continuous data from sources such as Apache Kafka, Online Archive moves aging data to low-cost object storage, and Data Federation queries data across clusters and cloud storage. The service is built for the operational data behind live applications, which distinguishes it from analytics platforms such as Snowflake and Databricks.
Atlas is MongoDB Inc.’s largest business, accounting for 73% of the company’s revenue in fiscal 2026. Developers can start on a permanent free tier and move to Flex or dedicated clusters as usage grows. Since acquiring Voyage AI in 2025, MongoDB has built Voyage embedding and reranking models into the service, has introduced automatic embedding generation inside Vector Search and, in August 2026, launched a managed Model Context Protocol server that lets AI coding agents work with live Atlas data.
Key features of MongoDB Atlas
- 01
Automated cluster operations
Atlas handles provisioning, replication setup, patching and version upgrades without taking clusters offline. Auto-scaling can raise or lower cluster tiers and expand storage as workloads change, and a Performance Advisor recommends indexes for slow queries.
- 02
Multi-region and multi-cloud clusters
Clusters can run in more than 125 regions on AWS, Google Cloud and Azure, and one cluster can spread its nodes across regions or providers. Automatic failover between them underpins an uptime commitment of up to 99.995%.
- 03
Backups and point-in-time recovery
Dedicated clusters take scheduled cloud backups and can keep continuous backups, so administrators can restore data to a specific moment, for example just before a faulty deployment corrupted records.
- 04
MongoDB Search
Full-text indexes built on Apache Lucene support relevance ranking, autocomplete, fuzzy matching and faceted results. Queries run through a $search stage in the aggregation pipeline, and dedicated search nodes can separate search load from the database.
- 05
Vector Search
Embeddings are stored alongside ordinary document fields and indexed for approximate nearest-neighbor search, which can be combined with keyword search and metadata filters. MongoDB has also introduced automatic embedding generation with Voyage AI models.
- 06
Atlas Stream Processing
Stream processors read continuous data from Apache Kafka and MongoDB change streams, transform and aggregate it over time windows using the database’s aggregation syntax, and write results to Atlas collections or back to Kafka.
- 07
Online Archive and Data Federation
Online Archive moves older documents from a cluster to cloud object storage on a schedule while keeping them queryable, and Data Federation lets a single query read from Atlas clusters and files in storage services such as Amazon S3.
- 08
Security and compliance controls
Data is encrypted in transit and at rest, customers can manage their own encryption keys, and private endpoints such as AWS PrivateLink keep traffic off the public internet. Atlas also provides database auditing, role-based access control and Queryable Encryption.
Who uses MongoDB Atlas?
- Startups and individual developers prototype on the free tier and move to Flex or dedicated clusters as traffic grows, without changing application code.
- Enterprises migrate self-managed MongoDB deployments to Atlas so that patching, backups, monitoring and scaling are handled by MongoDB.
- Global consumer apps and financial firms use multi-region clusters to keep response times low and to store personal data in the countries where rules require it.
- Teams building chat assistants and retrieval-augmented generation apps store documents, embeddings and operational records together and query them with Vector Search.
- Organizations with multicloud strategies run clusters across AWS, Google Cloud and Azure to reduce dependence on a single provider or to sit near other workloads.
History of MongoDB Atlas
MongoDB unveiled Atlas at its MongoDB World conference in New York on June 28, 2016, first on Amazon Web Services, and extended it to Microsoft Azure and Google Cloud in June 2017. Other companies already sold hosted MongoDB, and in 2018 MongoDB bought one of them, mLab, and moved its customers onto Atlas.
The service then broadened. Atlas Search, a Lucene-based feature now sold as MongoDB Search, arrived in 2020 along with Online Archive, the same year clusters could first span several clouds. Synchronization technology from Realm, acquired in 2019, became Atlas Device Sync for mobile apps, and Vector Search followed in 2023. MongoDB later pared back parts of the platform: Device Sync and the Atlas Data API reached end of life on September 30, 2025, and the Flex tier, generally available since February 2025, replaced serverless instances and small shared clusters, which Atlas stopped supporting in January 2026.
Atlas’s share of MongoDB revenue climbed from 66% in fiscal 2024 to 73% in fiscal 2026. The service marked its 10th anniversary in June 2026, and in August 2026 MongoDB launched a managed MCP server that connects AI coding agents to Atlas data.
- 2016
MongoDB launches Atlas on Amazon Web Services on June 28.
- 2017
Atlas becomes available on Microsoft Azure and Google Cloud.
- 2018
MongoDB acquires database hosting provider mLab and begins moving its customers to Atlas.
- 2020
Atlas adds multi-cloud clusters, Atlas Search and Online Archive.
- 2023
MongoDB releases Atlas Vector Search for AI applications.
- 2025
The Flex tier becomes generally available, and Device Sync and the Data API reach end of life.
- 2026
Atlas retires serverless instances in January and adds a managed MCP server for AI coding agents in August.
MongoDB Atlas pricing
Pricing modelUsage-based
Free clusters include 512 MB of storage; Flex clusters are billed by usage up to a monthly cap, dedicated clusters hourly by size, cloud and region, and add-on services separately.
Prices change often. Check the official MongoDB Atlas website for current plans.
MongoDB Atlas alternatives
Profiled on SoftwareLore:
Also considerAmazon DocumentDB, Azure Cosmos DB, Google Cloud Firestore, Couchbase Capella and Amazon DynamoDB
MongoDB Atlas: frequently asked questions
Who makes MongoDB Atlas?
MongoDB Atlas is made by MongoDB Inc., the New York company that develops the MongoDB database and trades on the Nasdaq under the ticker MDB. MongoDB launched Atlas in 2016, and the service has become its largest business, accounting for 73% of the company’s revenue in the fiscal year ended January 31, 2026.
What is MongoDB Atlas used for?
MongoDB Atlas is used to run MongoDB databases in the cloud without managing servers. Companies use it as the operational database for web, mobile and enterprise applications, and increasingly for AI applications that need vector search. It also covers full-text search, stream processing, data archiving and federated queries across clusters and cloud storage.
Is MongoDB Atlas free?
MongoDB Atlas has a free tier: a cluster with 512 MB of storage on shared hardware, intended for learning and small projects. Production use is paid. Flex clusters are billed by usage with a monthly cap, dedicated clusters are billed hourly by size, cloud provider and region, and services such as search nodes and stream processing cost extra.
When was MongoDB Atlas released?
MongoDB Atlas was launched on June 28, 2016, initially on Amazon Web Services. MongoDB added Microsoft Azure and Google Cloud in June 2017 and began offering clusters that span more than one cloud provider in 2020. The service marked its 10th anniversary in June 2026.
Which clouds does MongoDB Atlas run on?
MongoDB Atlas runs on Amazon Web Services, Google Cloud and Microsoft Azure, in more than 125 regions in total. Customers pick a provider and region for each cluster, and multi-cloud clusters can distribute nodes across two or three providers so that the database keeps running if one provider’s region fails.
What is the difference between MongoDB and MongoDB Atlas?
MongoDB is the database software itself, which organizations can download and run on their own servers. MongoDB Atlas is MongoDB Inc.’s managed cloud service that runs the same database for customers, taking care of setup, patching, backups and scaling, and it adds cloud-only services such as Atlas Stream Processing and Data Federation.
Sources
- MongoDB launches Atlas, its new database-as-a-service offering, TechCrunch techcrunch.com
- MongoDB pricing mongodb.com
- 10 years of MongoDB Atlas, MongoDB blog mongodb.com
- Migrate to Flex clusters, MongoDB Atlas documentation mongodb.com
- MongoDB, Inc. second quarter fiscal 2027 results sec.gov
Last reviewed . Spotted an error? Send a correction.