MongoDB makes and sells a general-purpose database platform built on a flexible, JSON-based document architecture. Unlike traditional relational databases that store data in rigid rows and columns, MongoDB lets developers store data in documents that naturally reflect how modern applications are structured — nested relationships, variable fields, and mixed data types. This makes it faster to build and iterate on applications, and every software application needs a database, so MongoDB's customers span virtually every industry. MongoDB sells through a developer-led, bottom-up motion — with free tiers and self-serve for smaller customers — alongside a direct enterprise sales force for larger accounts. Atlas, MongoDB's fully managed cloud-hosted database-as-a-service, accounts for roughly 73% of revenue. Atlas is consumption-based, meaning customers pay based on data stored and operations run, so revenue scales with customer application usage. Atlas runs on AWS, Google Cloud, and Azure, and includes integrated capabilities like full-text search, vector search, and stream processing. MongoDB Enterprise Advanced, roughly 20% of revenue, is a self-managed commercial license for customers running MongoDB on their own infrastructure, sold primarily as multi-year subscriptions and recognized upfront, which creates revenue volatility quarter to quarter. MongoDB's gross margins are in the mid-70s%, reflecting the software nature of the business. MongoDB positions itself as a natural foundation for AI applications, and its February 2025 acquisition of Voyage AI added embedding and reranking models to the platform, enabling MongoDB to serve as the operational data store, vector database, and embedding layer for AI applications in a single platform.