What is KRINT?

KRINT is a self-hosted platform for provisioning and operating databases. Pick an engine, click Launch, and you get a containerised instance with credentials, a host port, and a connection string already in hand.

• Provision in one click. 16 engines (Postgres, MySQL, Mongo, Redis, and more), each with opt-in plugins like pgvector and PostGIS.
• Browse and query. Spreadsheet-style in-cell row editing and an ad-hoc SQL console, in the browser.
• Container console. Live log tailing and an interactive shell into any provisioned instance.
• Back up and upgrade. Manual or cron-scheduled dumps; restore or upgrade the engine version in place.
• Manage users and access. Create logins, reset passwords, grant per-database access.
• Distribute across nodes. Provision databases onto remote Docker hosts over a single connection.
• Bring your own auth. OIDC with any provider (Pocket ID, Authentik, Auth0…), or the bundled Keycloak.

There is no limit to the number of databases or nodes you can run.

Architecture

KRINT runs from one image in one of two roles.

Role	What it does
Control plane (default)	The full app: UI, API, metadata database, and the secrets vault. All user interaction flows through it.
Node (Krint__Role=node)	A stripped, stateless worker that runs database containers on its own host and dials out to the control plane over one SignalR connection. Owns no state.

Each database is provisioned as an isolated sibling container on the Docker host - not a child of KRINT - which is why KRINT mounts the Docker socket.

Distributions

KRINT ships two ways from the same codebase.

Distribution	Metadata DB	Auth	Use case
Docker image	Postgres or SQLite	OIDC (bring-your-own or bundled Keycloak)	Servers, multi-user, always-on
Desktop app	SQLite	Built-in, zero-config	A single user on their own machine

Get started

• Self-hosting - run the image with docker compose.
• Desktop app - the single-user build.
• Developer setup - local dev with dotnet run + Bun.