Getting Started with TestMu AI MCP Server

TestMu AI MCP Server is a remote MCP server that connects MCP-compatible AI clients to your TestMu AI testing environment. It gives you access to tools for different TestMu AI products - HyperExecute, Automation, SmartUI, and Accessibility - which you can invoke through natural language prompts in your IDE or AI client.

Available Tools

---

TestMu AI MCP Server provides four tools, each covering a different area of testing:

1. HyperExecute Tool: AI-native test orchestration. Auto-generates test runner commands and HyperExecute YAML, monitors jobs, and surfaces session details.
2. Automation Tool: Triage and debug automation test failures. Pulls execution data, command/network/console logs, and helps identify root cause directly inside your IDE.
3. SmartUI Tool: Debug visual regressions. Returns natural-language summaries of pixel, layout, DOM, and perceptual differences from any SmartUI comparison run.
4. Accessibility Tool: Run accessibility checks on hosted URLs or local React apps and get detailed remediation guidance.

Connecting to TestMu AI MCP Server

---

TestMu AI MCP Server works with any MCP-compatible AI client. Pick your client below, paste the configuration, and complete the OAuth flow with your TestMu AI account.

• Server URL: https://mcp.lambdatest.com/mcp
• Authentication: OAuth (you will be redirected to testmuai.com on first use)

Cursor

Use the one-click install button to add the server to Cursor:

Or configure manually:

1. Open Cursor Settings > Tools & MCP and click Add new MCP server.
2. Paste the configuration:

{
  "mcpServers": {
    "mcp-lambdatest": {
      "url": "https://mcp.lambdatest.com/mcp"
    }
  }
}

3. Save and restart Cursor. Complete the OAuth flow when prompted.

Project-level configuration: Create .cursor/mcp.json in your project root to share with your team.

Claude Code

Run this in your terminal:

claude mcp add --transport http mcp-lambdatest https://mcp.lambdatest.com/mcp

Then run /mcp inside Claude Code, select mcp-lambdatest, and complete the OAuth flow.

Scope flags: --scope local (default, current project), --scope project (shared via .mcp.json), --scope user (all projects).

Claude Desktop

1. Open Settings > Connectors.
2. Click Add Custom Connector and enter the URL:

https://mcp.lambdatest.com/mcp

3. Complete the OAuth flow to connect your TestMu AI account.

Custom connectors in Claude Desktop are configured through Settings > Connectors, not the claude_desktop_config.json file. Custom connectors are available on Pro, Max, Team, and Enterprise plans.

Claude.ai (Web)

1. Go to claude.ai/settings/connectors.
2. Click Add custom connector and enter:
   • Name: TestMu AI
   • URL: https://mcp.lambdatest.com/mcp
3. Click Add, then Connect and complete the OAuth flow.
4. To use the connector in a chat, click the + button in the chat input and enable TestMu AI under Connectors.

GitHub Copilot (VS Code)

1. Open the Command Palette (Cmd+Shift+P / Ctrl+Shift+P) and run MCP: Add Server, choosing Workspace or Global. Or create .vscode/mcp.json manually:

{
  "servers": {
    "mcp-lambdatest": {
      "type": "http",
      "url": "https://mcp.lambdatest.com/mcp"
    }
  }
}

2. When prompted, confirm that you trust the server.
3. Complete the OAuth flow when first invoked. Use MCP: List Servers to manage installed servers.

Antigravity

1. In the Agent panel, click the ... menu, then MCP Servers, then Manage MCP Servers.
2. Click View raw config and add the configuration to mcp_config.json:

{
  "mcpServers": {
    "mcp-lambdatest": {
      "serverUrl": "https://mcp.lambdatest.com/mcp"
    }
  }
}

3. Save and refresh the MCP servers panel. Complete the OAuth flow when prompted.

note

Antigravity has a global limit of around 100 active MCP tools across all servers. If you are hitting that limit, disable tools you do not need from the MCP settings panel.

OpenAI Codex CLI

Run this in your terminal:

codex mcp add mcp-lambdatest --url https://mcp.lambdatest.com/mcp

Then start an OAuth login:

codex mcp login mcp-lambdatest

Verify with codex mcp list.

Manual config: Add the following to ~/.codex/config.toml:

[mcp_servers.mcp-lambdatest]
url = "https://mcp.lambdatest.com/mcp"

Cline (VS Code)

1. Open the Cline panel in VS Code, then MCP Servers > Configure MCP Servers.
2. Paste the configuration:

{
  "mcpServers": {
    "mcp-lambdatest": {
      "disabled": false,
      "timeout": 60,
      "command": "npx",
      "args": ["mcp-remote@latest", "https://mcp.lambdatest.com/mcp"],
      "alwaysAllow": []
    }
  }
}

3. Click Authenticate and complete the OAuth flow.

Continue

1. Create .continue/mcpServers/testmu-ai.yaml in your workspace and add:

name: TestMu AI mcpServer
version: 0.0.1
schema: v1
mcpServers:
  - name: TestMu AI
    type: streamable-http
    url: https://mcp.lambdatest.com/mcp

2. Switch to Agent Mode in Continue and complete the OAuth flow when prompted.

note

MCP servers in Continue work in Agent Mode only.

Other Tools

If your MCP client is not listed above, use one of these connection methods:

Transport	URL	Notes
Streamable HTTP (recommended)	https://mcp.lambdatest.com/mcp	Modern transport, widely supported
STDIO (via mcp-remote)	See below	Fallback for clients that only support stdio

Universal STDIO configuration for any client that accepts the standard mcpServers format:

{
  "mcpServers": {
    "mcp-lambdatest": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://mcp.lambdatest.com/mcp"]
    }
  }
}

A Typical Workflow

---

Once connected, open your project in your AI-enabled IDE and use natural language to invoke tools:

1. Open your project in your AI-enabled IDE.
2. Open the chat panel of your MCP client.
3. Type: Analyze the project and create a TestMu AI YAML file using the TestMu AI MCP Server.

Your client will:

• Analyze your project structure
• Identify your testing framework and test files
• Generate appropriate test runner commands
• Create a complete TestMu AI YAML configuration tailored to your project
• Save the YAML file to your project directory

You can also ask questions like:

• "What does failFast in HyperExecute mean?"
• "Triage the last failed test on TestID XXXXX."
• "Run a SmartUI analysis on comparisonId YYYYY."
• "Run an accessibility report on https://example.com."

Troubleshooting

---

OAuth flow does not open or fails

• Make sure pop-ups are not blocked in your default browser.
• If you have previously connected and the connection has expired, look for a Disconnect or Clear authentication option in your client's MCP settings, then reconnect.
• Verify your TestMu AI account has the necessary permissions for the workspace you are trying to access.

Tools not appearing after configuration

• Fully restart your client (close and reopen, not just reload the window).
• For Cursor, Claude Desktop, and Antigravity, this means quitting from the Dock or Taskbar.
• For VS Code, Cline, Continue, and Zed, reload the window via the Command Palette.
• For Continue, confirm you are in Agent Mode.
• For Antigravity, check that you have not exceeded the active tool limit across all MCP servers.
• Verify the configuration JSON, TOML, or YAML is valid (no trailing commas, matching braces, correct indentation).

Command not found: npx errors

This usually means your MCP client cannot find Node.js on its PATH. Use the full path to npx:

which npx   # macOS/Linux
where npx   # Windows

Then replace "command": "npx" in your config with the full path returned (for example, /usr/local/bin/npx).