What tech stack does Notion use?
Notion's product is a React and TypeScript front end backed by Node.js services and a heavily sharded PostgreSQL database fleet running on AWS. The detail below is detected from public sources — Notion's own engineering blog, StackShare, and job postings — so it reflects what Notion builds with rather than a vendor list it has confirmed.
- Frontend
- React · TypeScript
- Backend
- Node.js (Express)
- Data
- PostgreSQL (sharded)
- Cloud
- AWS
- Orchestration
- Kubernetes
- CI/CD
- Buildkite
Detected technologies
What Notion builds with, grouped by layer (from Notion's engineering blog, StackShare, and job posts).
- React· Frontend
- TypeScript· Frontend
- Redux· Frontend
- Node.js· Backend
- Express· Backend
- PostgreSQL· Data
- PgBouncer· Data
- AWS· Infrastructure
- Kubernetes· Infrastructure
- Buildkite· DevOps
What does Notion use on the backend and data layer?
Notion's back end runs on Node.js with Express, and its most distinctive engineering choice is its data layer: a single PostgreSQL database that Notion famously sharded into a fleet to handle billions of blocks, fronted by PgBouncer to pool connections and keep the database stable under high concurrency. It runs on AWS (EC2, S3, RDS) with Docker and Kubernetes for orchestration. The block-based data model — every page is a tree of blocks — is what makes the storage problem hard and the sharding work notable.
What does Notion use on the frontend?
Notion's interface is built with React and TypeScript, using Redux for state management, and is engineered around the same block model that defines the product. The web app is the primary surface, with the editor's responsiveness a constant engineering focus given how much structured content a single page can hold.
What Notion's stack means if you sell to them
Notion's heavy PostgreSQL, AWS, and Node footprint maps cleanly to database, observability, infrastructure, and developer-tooling pitches, and its public writing about scaling Postgres signals a team that engineers deeply and prefers to solve hard problems in-house. Integration and augmentation tend to land better than wholesale replacement. Because stack data is detected from job posts and tools like StackShare, treat it as directional and confirm specifics during discovery.
As of June 2026.Sources:Notion Engineering — sharding PostgresStackShare — Notion
Notion — frequently asked questions
