Command Line Interface (CLI)

Xata provides a CLI as a convenience for developers. It is intended to provide feature parity with Xata's Web UI (User Interface) to enable developers to build applications at rapid pace with as little friction as possible.


Our CLI is distributed as an npm package and can be installed using npm if you've got Node.js installed. We will publish a more universal binary for Windows, macOS and Linux in the near future.

Installing the Xata CLI

Installing with npm

To install the Xata CLI globally, run the following command:

npm i --location=global

This will install the Xata CLI. We recommend installing it globally because it becomes much more convenient to work with: you can invoke a xata command from anywhere, instead of the more convoluted variant with a project-scoped installation (npm run xata), or an npx-based installation (npx These alternative approaches also work according to your preference, but we recommend a global installation.

If you are installing the Xata CLI on Windows using Powershell you may need to wrap this part of the install command in quotes: ''


To use the Xata CLI, you'll have to authenticate with it to access your workspaces and databases. You can do this in two ways: scoped to a project (project mode), or globally (global mode).

Global Mode

To authenticate globally, across your entire system, run xata auth login. This will give you two options:

  • Create a new API Key. This will open your browser and, when you're logged in to Xata, allow you to create a new API key for use with the CLI.
  • Use an existing API Key. This will prompt for an existing key you have, which you can paste into your terminal. We recommend isolating keys per use-case, so creating a new one might be a good idea instead.

Once you supply an API key one way or another, your CLI will be configured globally: we'll store your API key in ~/.config/xata/credentials.

From now on, when you use the Xata CLI outside of a project (i.e, globally) for database operations, you'll have to explicitly use a flag to indicate which database you'd like to work with. You can do this using the --db [url] flag. If you omit this flag, you will be interactively prompted to choose a database.

Working with workspaces globally does not require any further flags.

Project Mode

To authenticate in a specific project, run xata init --db=[databaseUrl]. You can get your database URL from the workspace configuration page in the web UI. This will create some configuration files in your project that the CLI will reference when working with Xata. Here's what happens when you initialize a project:

  • If you don't have a workspace, you will be prompted to create one.
  • If you have workspaces, you'll be asked to choose one, or to create a new one.

A workspace is a logical grouping of databases, usually analogous to an organization or team, so this is the first step. Once you've chosen a workspace, you will be given the option to either create a new database, or use an existing one for your project. After choosing a workspace and a database, you're ready to go: the CLI will walk you through next steps. Specifically, it will:

  • Create a project configuration file in your current working directory (.xatarc).
  • Create or update your .env file to store a database-scoped API key. (At the time of writing these docs it'll actually store a personal API key, but this will change soon).
  • Ask you if you'd like to install the SDK and/or use the TypeScript/JavaScript code generator.

When you have a project set up, the Xata CLI will now be aware of your project's configuration, namely which workspace, database, and fallback branch you've chosen. It will know this information by reading it from a new set of files created in your current working directory: .xatarc and .env. Let's discuss these briefly.


This file contains information about your current Xata project, specifically its database URL, and preferred output path of the generated Xata client that you can use to interact with your database if you've chosen to use it. More about this in code generation.

We recommend checking this file in to your version control system (Git, SVN, etc.) so that you can easily share your project's configuration among your team.


This file contains sensitive information and secrets that ought not be committed to version control. It is recommended that you add this file to .gitignore, so that it is not accidentally committed to version control. The Xata CLI appends secrets to this file, namely your API key and fallback branch. More on branches in the git section.

Code Generation

We strongly recommend using the CLI to generate your Xata client for you. The Xata client is a class that you can instantiate in your project that will help you query Xata in a familiar way. This will help make your queries to Xata predictable and safe, while also providing autocomplete and safety.

Once you have a generated Xata client, you can import it into your project and use it. The SDK Playground in the Xata web interface can serve as a companion for you to build and test out your queries. Once you're happy with them, you copy and paste them into your project.

As your schema changes, running xata codegen will consistently refresh your auto-generated Xata client with the latest schema for your database branch.

Branches & Git Integration

Xata is a branchable database. Developers can checkout new branches of their database schemas, make changes safely, and merge them into their main branch. This triggers a zero-downtime migration.

Some commands are integrated with git. Specifically, those related to branching. Xata uses the following algorithm in development in order to calculate which xata branch needs to be used:

graph TD A[Get current git branch] -->B(Is there a mapped Xata branch <br>with the current git branch?) B --> |yes| C(Use it) B --> |no| E(Is there a Xata branch with <br>the name as the git branch?) E --> |yes| C E --> |no| G(Is there a fallback branch?) G --> |yes| C G --> |no| I(Use the default branch)

The current git branch is used in the first two cases in the algorithm:

  • If a git branch has been explicitly mapped to a Xata branch, the associated branch is used
  • Otherwise, if a Xata branch with the same name as the git branch exists, we use that.

If none of them apply, the fallback branch is used. This is a safety measure to avoid connecting to your production database. If you are developing a feature that doesn't require schema changes, you'll have a git branch, but not an associated Xata branch. In this case to prevent using the default Xata branch—which could be production—we allow specifying a XATA_FALLBACK_BRANCH in your local .env file. You can have a personal branch with your username for example.

If none of the above apply, the default Xata branch is used.

Editing Your Database Schema

You can edit the schema in the CLI with xata schema edit but you can also quickly jump to the Web UI with xata browse and edit the schema there.

If you edit the schema in the CLI, the code generator will be used if it's configured after applying the schema changes. If you choose to edit the schema in the web UI, remember to execute xata codegen when you are back to your code.

Next Steps

Now that your project is set up with the Xata CLI, and if you've opted to use code generation (recommended), we suggest exploring the TypeScript/JavaScript SDK section of the documentation to get up to speed using it.