What it is

Conar is an AI-powered open-source database management project built for PostgreSQL, MySQL, MSSQL, and Clickhouse, with support for other databases indicated as coming in the future. It lives in the TypeScript and AGPL-3.0 database management ecosystem, categorized under Infrastructure & Operations / Databases. It addresses the friction of working with multiple database systems while needing help writing and optimizing SQL queries. The README describes storing connections securely in the project’s cloud and asking AI to assist with SQL. It also describes encrypted connection strings and password protection, which aims to reduce the risk of handling database credentials and query logic manually across different engines.

Key capabilities

• Conar supports PostgreSQL, MySQL, MSSQL, and Clickhouse as current database targets, with SQLite and MongoDB listed as coming soon.
• Conar provides AI-powered SQL assistance for writing and optimizing queries, using the Vercel AI SDK with Anthropic.
• Conar allows users to change the AI model, which is listed under AI-Powered Features.
• Conar stores connection strings with encryption and supports password protection, as described in the Secure & Open section.
• Conar uses Drizzle ORM for database migrations and schema preparation, as shown in the development setup.
• Conar includes an optional agent-browser workflow that can drive the running app for agents and manual debugging.

Who uses it and how

• Developers can use it as a local or cloud-assisted database interface for PostgreSQL, MySQL, MSSQL, or Clickhouse connections and AI-assisted SQL help.
• Teams can run it from source by installing dependencies with pnpm, starting Postgres and Redis with Docker or hosted instances, running migrations, and launching development servers.
• Agents can interact with the running application through agent-browser commands such as open, snapshot, click, and console, or through another browser automation tool.
• Users can validate behavior through unit tests and end-to-end tests, after starting the required test server.

Getting started

The local setup uses pnpm install, an optional global npm install -g agent-browser, Docker for Postgres and Redis or hosted instances, and then pnpm run drizzle:migrate and pnpm run dev. The README also mentions a hosted cloud option for storing connections and an agent-browser package for browser automation.

When to use it — and when not to

Conar fits users who want an open-source, AI-assisted database interface and can manage Postgres, Redis, migrations, and local development services themselves. It may not fit users who want to avoid setup, because the README describes source-based development setup and backing services that must be reachable. Its zero-year repository age, zero listed contributors, and 46 open issues indicate that the project is still forming and should be evaluated carefully before production use.