# e2e-runner MCP server

> An AI-native end-to-end test runner that uses JSON-driven test definitions to write, run, and debug web application tests via MCP tools.

- Repository: https://github.com/fastslack/mtw-e2e-runner
- Page: https://hostedmcp.org/servers/fastslack/mtw-e2e-runner
- Transport: stdio
- Stars: 7
- License: Apache-2.0
- Latest version: 1.3.0

## Install

### Claude Code

```sh
claude mcp add mtw-e2e-runner -- npx -y @matware/e2e-runner
```

### Cursor (`~/.cursor/mcp.json`)

```json
{
  "mcpServers": {
    "mtw-e2e-runner": {
      "command": "npx",
      "args": [
        "-y",
        "@matware/e2e-runner"
      ]
    }
  }
}
```

### VS Code (`.vscode/mcp.json`)

```json
{
  "servers": {
    "mtw-e2e-runner": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@matware/e2e-runner"
      ]
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"mtw-e2e-runner","type":"stdio","command":"npx","args":["-y","@matware/e2e-runner"]}'
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "mtw-e2e-runner": {
      "command": "npx",
      "args": [
        "-y",
        "@matware/e2e-runner"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add mtw-e2e-runner -- npx -y @matware/e2e-runner
```

### Codex config.toml (`~/.codex/config.toml`)

```toml
[mcp_servers.mtw-e2e-runner]
command = "npx"
args = ["-y", "@matware/e2e-runner"]
```

### Gemini CLI

```sh
gemini mcp add -s user mtw-e2e-runner npx -- -y @matware/e2e-runner
```

### Gemini settings.json (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "mtw-e2e-runner": {
      "command": "npx",
      "args": [
        "-y",
        "@matware/e2e-runner"
      ]
    }
  }
}
```

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `CHROME_POOL_URL` | no | no |  |

## FAQ

### Do I need to write code for my tests?

No, tests are written in plain JSON files, and you can even use Claude Code to generate them through conversation.

### Do I need Docker to run tests?

No, you can use your existing Chrome browser or the Obscura binary without Docker. Docker is only required if you want to use the parallel Chrome pool.

### How does the AI integration work?

The runner provides 17 MCP tools that allow AI agents like Claude Code to create, execute, and debug tests natively.
