PlugWiki
Auth0 MCP Server logo

Auth0 MCP Server: Setup for Claude, Cursor and Other Apps

OfficialLocal
Getnpm@auth0/auth0-mcp-serverauth0/auth0-mcp-server

Lets Claude, Cursor and other assistants manage your Auth0 apps, APIs, login actions, forms and logs.

This is Auth0's own server. It runs on your computer and talks to your Auth0 tenant, so you can ask your assistant to list apps and APIs, read logs, create apps, and edit actions and forms. You sign in once through your browser with your Auth0 account, and the server keeps the login in your system keychain. Add the read-only switch if you only want it to look.

It is a beta release (version 0.1.0-beta.19). It can change live settings, including deploying login code, so check what it plans before you approve each step.

Good for

  • Developers adding Auth0 to a new app who want the assistant to create the application and wire up the SDK
  • Admins who want to check recent failed logins and tenant settings by asking in plain words
  • Teams that keep Auth0 actions and forms in the dashboard and want help drafting them

Works in

12 should work

Tap an app to see how to set it up and any known problems.

Install Auth0 MCP

Claude Code

  1. 1

    Copy this prompt.

    Add the Auth0 MCP server to Claude Code. Run this command:
    
    npx @auth0/auth0-mcp-server init --client claude-code
    
    Then check that it loaded and tell me if I need to sign in or restart.
  2. 2

    Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js 18+, Auth0 account.

Read-only mode: Add --read-only after run in the server arguments. Only the list and get tools stay.

The guided setup signs you in through your browser and writes this client's config for you. If you add the config by hand instead, run init first so the server has a login. Requires Node.js 18 or later and an Auth0 account.

From the project README. Source.

Tools

  • Lists your Auth0 applications, or searches them by name

    Returns the applications in your tenant. You can ask for all of them or search by part of a name.

    Try: “Show me every Auth0 application with "api" in its name”

    Network · auth0_list_applications

  • Shows the settings of one Auth0 application

    Fetches one application by client ID or name, including its callback and logout URLs and type.

    Try: “What are the callback URLs for my Customer Portal app?”

    Network · auth0_get_application

  • Lists the APIs registered in your tenant

    Returns the APIs (Auth0 calls them resource servers) your tenant protects, with their identifiers.

    Try: “Which APIs are set up in my Auth0 tenant?”

    Network · auth0_list_resource_servers

  • Shows the settings and scopes of one API

    Fetches one API by ID or identifier, including the scopes it defines and its token settings.

    Try: “What scopes does my Payment API define?”

    Network · auth0_get_resource_server

  • Lists the login and flow actions in your tenant

    Returns the actions (small pieces of code that run during login and other flows) in your tenant.

    Try: “What actions are configured in my tenant?”

    Network · auth0_list_actions

  • Shows the code and status of one action

    Fetches one action, including its code, trigger and whether it is deployed.

    Try: “Show me the code for my Add Custom Claims action”

    Network · auth0_get_action

  • Lists the custom forms in your tenant

    Returns the forms (custom login, signup and similar pages) your tenant has.

    Try: “List the custom forms in my tenant”

    Network · auth0_list_forms

  • Shows the settings of one form

    Fetches one form by ID with its configuration.

    Try: “Show me the configuration of my Corporate Login form”

    Network · auth0_get_form

  • Searches the tenant log for sign-ins and other events

    Returns tenant log entries. You can filter, for example for failed logins in a time window.

    Try: “Find failed logins from the past 24 hours”

    Network · auth0_list_logs

  • Shows one log entry in full

    Fetches a single log entry by its ID so you can see what happened.

    Try: “Get more detail on log entry abc123”

    Network · auth0_get_log

  • Creates a new Auth0 application

    Creates an application of the type you pick (single-page, native, machine-to-machine and so on). The client secret is hidden in the reply and you view it in the dashboard.

    Try: “Create a single-page application called Analytics Dashboard”

    Write · auth0_create_application

  • Changes the settings of an Auth0 application

    Updates an existing application, for example its callback URLs, logout URLs or metadata. The tool marks it as able to overwrite existing values.

    Try: “Add https://staging.example.com/callback to the callback URLs of my Web App”

    Write · auth0_update_application

Read from the source code

Permissions

medium risk · 5 things to watch

Medium risk

With the default setup the assistant can change your live Auth0 tenant, including deploying code that runs at login, and none of the 23 tools deletes anything. Start with the read-only switch and grant only the permissions you need when you sign in.

What it can do

Watch out for

  • It can deploy code that runs at login

    The assistant can write an action and make it live. Ask for the code first, read it, and deploy it yourself, or use read-only mode. github.com

  • It can change apps, APIs and who can call them

    It can edit callback URLs and token settings, and give an app access to an API. Review each change, and sign in with only the scopes you need. github.com

  • It writes app secrets to a .env file

    Two tools save the client secret into your project folder. Check the file is in .gitignore, which the tool adds for you. github.com

  • Usage stats go to a third party by default

    Anonymized data such as version, operating system and timestamp is sent. Set AUTH0_MCP_ANALYTICS to false to stop it. github.com

  • Your login stays until you log out

    The sign-in token lives in the keychain until it expires. Run the logout command when you are done or switch tenants. github.com

Security advisories

No published security advisories as of Nov 1, 2026.

Technical details
Turning on read-onlyAdd --read-only to the run arguments, or run init with --read-only. Only the 10 list and get tools stay.
Talks toYour Auth0 tenant (Management API). cdn.auth0.com (framework quickstart guides). heapanalytics.com (anonymized usage stats, can be turned off)
Sign-in and keysLogin tokens are kept in the system keychain; the logout command removes them.
Scopes
  • read:clients

  • create:clients

  • update:clients

  • read:resource_servers

  • create:resource_servers

  • update:resource_servers

  • read:actions

  • create:actions

  • update:actions

  • read:logs

  • read:forms

  • create:forms

  • update:forms

Questions

Is the Auth0 MCP server official?

Yes. It is published by Auth0 in its own GitHub organization under the MIT license. It is still a beta release.

Does it need an Auth0 account and an API key?

It needs an Auth0 account but no API key. Running the init command opens your browser so you can sign in and pick the tenant and permissions. Private cloud tenants cannot use that sign-in and use a machine-to-machine app with a client ID and secret instead.

Which apps does it work with?

The init command can set up Claude Desktop, Claude Code, Cursor, VS Code, Windsurf and Gemini CLI for you. Other clients, including Codex, use a manual config. Any app that can run a local server works.

Can I make it read-only?

Yes. Add the read-only option when you run init or when the server starts. That leaves only the 10 list and get tools, even if you also pick other tools. You can instead limit tools with name patterns.

Can it delete things in my tenant?

No. None of the 23 tools deletes anything. It can still overwrite settings, change an action's code and deploy it, so review changes before you approve them.

Does it run locally or in the cloud?

It runs on your computer and calls your Auth0 tenant directly. It also sends anonymized usage stats to Heap Analytics unless you set AUTH0_MCP_ANALYTICS to false.

Why does it say I need to sign in again?

Your login token expired, you logged out, or you want a different tenant. Run the init command again. The run command checks the token before it starts.

Does it use up my Auth0 rate limits?

Yes. Every tool call is a Management API call and counts against the limits of your Auth0 plan.

Information

updated Nov 2, 2026

Publisher
Official
Package
@auth0/auth0-mcp-server
Registry
npm
Transport
stdio
Auth
OAuth
Version
0.1.0-beta.10
License
MIT
Last commit
Oct 20, 2026
Open issues
8
Listed in
Official MCP registry

GitHubnpmOfficial MCP RegistryChangelogIssues

Sources

Docs, issue trackers and discussions this page is built from. Issues, quotes and findings above link to their own source.

Show all 2
  1. Auth0 MCP Server README · github.com
  2. Tool registrations at the v0.1.0-beta.19 release · github.com

Facts from the official MCP registry, GitHub, npm, PyPI and the Docker catalog. Updated Nov 2, 2026.