---
title: "UXArmy MCP Server"
canonical: "https://help.uxarmy.com/space/general/735870977/UXArmy%20MCP%20Server"
format: markdown
---
Connect UXArmy to supported AI assistants like Claude, ChatGPT, Cursor, etc. and work with your research workspace using natural-language prompts.

## **What is UXArmy MCP?**

UXArmy MCP (Model Context Protocol) connects your UXArmy workspace with supported AI assistants, allowing you to interact with your research without having to navigate through the UXArmy platform for every task.

Once connected, you can ask your AI assistant to find projects and studies, review study setup and research results, explore participant responses, and retrieve information from your UXArmy workspace.

You can also use supported MCP actions to create surveys, update studies and perform other research tasks directly through your AI assistant.

For example, you can ask: “Show me my UXArmy projects.” or: “Create a survey about checkout frustration with six questions.”

UXArmy MCP is constantly being enhanced, so the available capabilities and supported AI tools may change as we continue improving the experience.

## **Set up UXArmy MCP in Claude**

Before you begin, make sure you are **signed in to your UXArmy account in your browser**. The MCP connection uses the UXArmy account that you approve during authorisation.

### **If UXArmy is already available as a connector**

If UXArmy has already been added to your Claude workspace:

1. Open **Claude**.
2. Go to **Settings → Connectors**.
3. Find **UXArmy** in the connector list.
4. Select **Connect**.
5. Approve the access request.
6. Once connected, start a new Claude conversation and test the connection.

The UXArmy connector uses the following MCP endpoint:

`https://app.uxarmy.com/rest/mcp`

## **Add UXArmy as a custom connector**

If UXArmy isn't already available in your connector list:

### **1. Open Claude's connector settings**

Go to:

**Settings → Connectors → Add custom connector**

### **2. Enter the UXArmy connector details**

**Name:** UXArmy

**URL:**

`https://app.uxarmy.com/rest/mcp`

Leave the other settings unchanged and continue.

![Add UXArmy as a custom connector.png](media://0b32a7f3-4185-4ccc-ad28-01a7f4eb00d8)

*Add UXArmy as a custom connector*


### **3. Authorise UXArmy**

Claude will open a UXArmy authorisation page.

Review the account being connected and the permissions requested. Depending on the connection, the permissions may include access to:

- Projects
- Studies
- Results
- Profile
- Workspaces
- Subscription information

If the connection includes build permissions, it may also allow Claude to create, edit or delete supported projects and studies. Select **Allow** to approve the connection.

![Add Authentication.png](media://92032ecb-37df-4daa-992c-caf875ea7b98)

*Add Authentication*

## **Verify your connection**

Once connected, start a new Claude conversation and try: List my UXArmy projects.

If Claude returns your UXArmy projects, your connection is working.

You can also try:

What UXArmy workspaces do I have access to? Or: Show me my recent UXArmy studies.

## **What can you do with UXArmy MCP?**

Once connected, you can use natural-language prompts to access information and perform supported actions in UXArmy.

## **View your projects**

You can ask Claude to:

- List your projects
- See how many studies each project contains
- View a specific project and its studies
- Search for projects

For example:

*List my projects and how many studies are in each.*

*Show me the studies in the Onboarding project.*

![View Your projects-png.png](media://4c2cee41-9b94-4d5a-86d4-bc8f2fb8a1f2)

*View Your projects*

## **Review your studies**

You can ask about:

- Study type
- Study status
- Workspace
- Project
- Creation date
- Responses received
- Study setup
- Participant targeting and settings

For example:

*Which studies are active right now?*

*Find the study with "checkout" in the name.*

*Show me the full setup of the pricing card sort.*

## **Review Tree Testing and Card Sorting studies**

UXArmy MCP can also retrieve information from Tree Testing and Card Sorting studies.

For Card Sorting, it can access information such as:

- Cards
- Categories
- Card Sorting type
- Whether explanations are requested
- Instructions and thank-you screens
- Language
- Response target
- Screener and survey questions

## **Review study results**

### **Tree Testing**

You can ask Claude about:

- Success rate
- Directness rate
- Task outcomes
- First nodes clicked
- Where participants stopped
- Correct answers
- Participant paths
- Completion time

Example:

*What are the success and directness rates for the navigation tree test?*

### **Card Sorting**

You can ask:

*Where did people actually put each card?*

*Which cards were the most disagreed-about?*

*Which categories did participants create themselves?*

UXArmy MCP can retrieve card/category results and participant-created categories.

## **Review individual participant responses**

You can also ask Claude to look into an individual response.

For example:

*Open response 4 from the checkout study.*

*Which task did that participant get stuck on?*

*What did they say out loud while doing it?*

For recorded studies, UXArmy MCP can provide the transcript, including timed participant utterances and an AI-generated summary.

## **Check your UXArmy account**

You can ask Claude about:

- Your profile
- Organisation
- Workspaces
- Workspace members and roles
- Subscription plan
- Seats
- Credits
- Credit usage
- Interview availability

For example:

*What plan are we on, and how many seats are left?*

*How many credits do we have?*

*What did we spend credits on this month?*

**Note:** Billing information, including plan, credits and usage, is available only to organisation owners, admins or billing admins.

## **Create a survey with UXArmy MCP**

UXArmy MCP can create a **draft survey** using Claude.

For example:

**Draft a 6-question survey on checkout frustration in the Onboarding project.**

You can specify different question types, including:

- Yes/No
- Multiple choice
- Open text
- Short answer
- Matrix
- Ranking
- Likert scale
- Net Promoter Score
- Notes

The survey is initially created as a **draft**, so it isn't visible to participants and doesn't spend credits.

You can then ask Claude to:

*Add an NPS question at the end.*

or:

*Make question 3 optional and shuffle its options.*

![Create a survey with UXArmy MCP.png](media://15900099-71f3-44cc-ae09-7d66d01299ea)

*Create a survey with UXArmy MCP*


![Open the survey link from Claude.png](media://6d6435ca-75b1-4c23-b6db-23e9901fd243)

*Open the Survey link from Claude in UXArmy*

## **Edit an existing study**

You can use Claude to make supported changes to existing studies, including:

- Rewording tasks/questions
- Adding or removing tasks
- Reordering tasks
- Renaming a study
- Changing welcome and thank-you screens
- Changing recruitment settings
- Updating redirects
- Updating demographics
- Changing task shuffling
- Updating SUS questions

For example:

*Reword task 2 and remove task 5 in the checkout study.*

![Ask Claude to edit your existing study.png](media://37f79592-8c1f-4121-91a4-db8a7f348980)

*Ask Claude to edit your existing study*

## **Launch and publish changes**

You can ask Claude to launch a draft or publish changes to an existing live study.

For example:

*Show me the questions, then launch it.*

or:

*Publish the changes I just made.*

However, **launching is a real-world action**. It can open a study to participants, use account credits and potentially send invitation emails. Claude should tell you what will happen before carrying out the action.

### **Important**

A new study does not automatically have participants assigned to it. You need to specify who should take the study before launching.

## **Connect UXArmy MCP to ChatGPT**

**ChatGPT Pro users** can connect MCPs with read/fetch permissions through Developer Mode. 

## **How to Connect UXArmy MCP**

### **1. Open ChatGPT Settings**

Go to:

**ChatGPT → Settings → Apps**

If you have access to Developer Mode, enable it from the available **Advanced Settings**.

![ChatGPT Apps Settings Interface.png](media://db7e8ade-df92-493a-a7dc-3c9f147e929e)

*ChatGPT Apps Settings Interface*

### **2. Create a Custom MCP App**

Select:

**Apps → Create**

Enter:

**Name:** `UXArmy`

**MCP endpoint:**

[https://app.uxarmy.com/rest/mcp](https://app.uxarmy.com/rest/mcp)

### **3. Configure Authentication**

Choose the authentication method provided for the UXArmy MCP server.

If ChatGPT displays the authentication option, select:

**Always required**

Then click **Connect/Reconnect** and sign in to your UXArmy account.

**Important:** Never enter your UXArmy password or access token directly into the MCP configuration or ChatGPT conversation.

![Creating a Custom MCP Connector.png](media://74dc73b7-6d79-4915-8804-547332ef8684)

*Creating a Custom MCP Connector*

### **4. Scan UXArmy Tools**

After authentication:

1. Click **Scan Tools**.
2. Wait for ChatGPT to retrieve the available UXArmy tools.
3. Review the tools.
4. Click **Create**.

OpenAI's current MCP setup flow uses the endpoint, authentication configuration, **Scan Tools**, and then **Create**.

### **5. Test the Connection**

Open a new ChatGPT chat and select **UXArmy** from the tools/apps menu.

Try:

*Show my UXArmy projects.*

If the connection is successful, ChatGPT will use the UXArmy MCP tools to retrieve the information.

## **Connect UXArmy MCP Server to Gemini CLI**

Connecting a customized Model Context Protocol (MCP) server for UXArmy (or any custom UX testing service) to Gemini depends on whether you are using **Gemini Enterprise (Business/Cloud UI)** or the local **Gemini CLI**.

Because MCP servers communicate via JSON-RPC, your custom UXArmy wrapper must first expose an **HTTP/SSE endpoint** (for remote server setups) or a **stdio process** (for local CLI setups).

## **Prerequisites**

Ensure you have Node.js version 20 or higher installed on your system.

Install the Gemini CLI globally by executing the following command in your terminal:

Bash

npm install -g @google/gemini-cli

## **1. Configure the UXArmy MCP Server**

You can register the UXArmy MCP server using either the terminal CLI command or by manually updating your configuration file.

### **Option A: Via Terminal Command (Recommended)**

Run the following command in your terminal to automatically update your global configuration:

Bash

gemini mcp add --transport http --scope user uxarmy [https://app.uxarmy.com/rest/mcp](https://app.uxarmy.com/rest/mcp)


### **Option B: Via Settings File**

Add the configuration snippet directly to your `settings.json` file located at:

- **macOS / Linux:** `~/.gemini/settings.json`
- **Windows:** `C:\Users\<username>\.gemini\settings.json`

JSON

{

  "mcpServers": {

    "uxarmy": {

      "httpUrl": "[https://app.uxarmy.com/rest/mcp](https://app.uxarmy.com/rest/mcp)"

    }

  }

}

## **2. Authenticate**

Launch the Gemini CLI by running:

Bash

gemini

Inside the active Gemini CLI session, run the authentication command:

Plaintext

/mcp auth uxarmy

- A browser window will open automatically.
- Sign in with your UXArmy credentials to authorize the connection.
- The access token will be securely saved and reconnected automatically for future sessions.

**Important:** Never input your UXArmy password, access token, or API keys directly into the CLI prompt or chat conversation.

## **3. Verify Connection**

Inside the Gemini CLI session, verify that the UXArmy server is linked and displaying its tools by running:

Plaintext

/mcp

Check the command output to ensure `uxarmy` appears under connected tools.

## **4. Test the Integration**

Test the connection by running a prompt directly in your Gemini CLI session:

Plaintext

Show my UXArmy projects.

If configured correctly, Gemini CLI will invoke the UXArmy MCP tools to retrieve and display your research workspace data.