Tabularis is an open-source, cross-platform desktop SQL workspace that integrates built-in AI agent support via an MCP server. It supports managing connections, running queries in SQL notebooks, and visualizing execution plans across multiple databases.
Project overview
It provides a built-in MCP server enabling AI agents to read schemas and execute queries, along with a plugin system that allows database drivers to be written in any language communicating over JSON-RPC.
Project type
MCP · Data Processing
Use cases
Data Analysis
Deployment
Refer to project documentation
License
Apache-2.0
Best for
Developers, data teams, and AI engineers who need a cross-platform desktop GUI for multi-database SQL querying and want to connect AI agents via MCP.
Users who want optional text-to-SQL capabilities and the flexibility to choose between external services like OpenAI or local models via Ollama.
Key capabilities
Manage connections, explore database schemas, and run SQL queries using an integrated Monaco editor.
Combine SQL and Markdown cells in a single document, featuring inline results, cross-cell variables, and charts.
Inspect database execution plans as navigable node graphs.
A built-in MCP server that allows AI agents to read database schemas and execute queries.
An optional text-to-SQL and query explanation feature powered by external or local AI models.
Build queries visually with drag-and-drop tables.
Extend database support via standalone executable plugins that communicate over JSON-RPC.
Interactive Entity-Relationship visualization with selective table diagram generation.
Limitations and risks
The project focuses on doing a few databases well rather than supporting dozens of drivers out-of-the-box, which may require users needing niche databases to write their own plugins.
Getting started
Installation is documented as easy, with an installer available via package managers such as winget, brew, and snap. An example first success path is running winget install Debba.Tabularis and then opening Tabularis.
Alternatives and comparisons
A free, cross-platform database client with a full-featured SQL workspace and an AI assistant that operates on the user's machine.
A local MCP server providing full-featured access to MongoDB databases using natural language via LLMs.
Provides database interaction with Snowflake via the Model Context Protocol for SQL querying and schema exploration.
GitHub project description: Open-source desktop SQL workspace for PostgreSQL, MySQL/MariaDB, SQLite and 13+ more databases like DuckDB, ClickHouse, Redis and Firestore. Built-in MCP server for Claude, Cursor…
README: Support for **MySQL/MariaDB**, **PostgreSQL** (with multi-schema support) and **SQLite**, with multi-database selection per connection.
README: Save, manage, and clone connection profiles, with optional secure password storage in the system **Keychain**.
README: AI text-to-SQL with **local models** (Ollama) | ✅ | Cloud-based AI assistant | ❌ | ❌
README: PostgreSQL, MySQL/MariaDB and SQLite ship built in. Everything else is a plugin