Using Sheeta AI

Projects

A project is a folder on your computer that groups related database work. Sheeta AI always works inside a project. Because a project is a plain folder, you can share it with your team through Git.

What a project contains

Sheeta AI keeps its own data in a .sheeta folder inside the project:

  • Connections in connections.json, with credentials in a separate .env file that Git ignores. See DB Connections.
  • Dashboards in canvases/.
  • Table and column descriptions in table-metadata/ and schema-metadata/.
  • AI chat history in ai-chat/ and AI permission rules in ai-permissions.json.

Your own files, such as SQL scripts, CSV files and notes, live in the project folder next to .sheeta. Local SQLite databases created in Sheeta AI are stored in a sqlite-db folder in the project.

Your first project

The first time you open Sheeta AI, it creates and opens a project folder named Sheeta in your Documents folder. After that, it reopens the most recent project on each launch.

Create or open a project

  1. Click the project name at the top of the left sidebar, or press Cmd+Opt+O.
  2. Choose Open Project. On the start screen, click Open folder or press Cmd+O.
  3. Select an existing folder, or create a new empty folder in the picker and select it.

The same menu lists Recent Projects (up to 10), with a search box and Remove from recent.

Opening an existing Sheeta AI project

Clone a project from Git

  1. Choose Clone from Git in the project menu, or click Clone repo on the start screen (Cmd+G).
  2. Enter the Repository URL and choose a Destination Folder. The URL must start with https://, http:// or git@.
  3. Click Clone Repository.

Cloning uses the git command on your computer, so Git must be installed. After cloning, you see the team's connections, files and dashboards. Enter your own database credentials in Connection Management, because they are not stored in Git.

Cloning a Sheeta AI project from Git

Open a project from the terminal

After installing the sheeta command, run sheeta . in a folder to open it as a project. See CLI and MCP.

Next steps