Zero-Config, Fully AI-Managed End-to-End Testing for all code gen platforms.
65 stars
TypeScript
Updated Oct 18, 2025
automation
browser
end-to-end-testing
web
Documentation
Debugg AI — MCP Server
AI-powered browser testing via the Model Context Protocol. Point it at any URL (or localhost) and describe what to test — an AI agent browses your app and returns pass/fail with screenshots.
Setup
Get an API key at debugg.ai, then add to your MCP client config:
json
{
"mcpServers": {
"debugg-ai": {
"command": "npx",
"args": ["-y", "@debugg-ai/debugg-ai-mcp"],
"env": {
"DEBUGGAI_API_KEY": "your_api_key_here"
}
}
}
}Or with Docker:
bash
docker run -i --rm --init -e DEBUGGAI_API_KEY=your_api_key quinnosha/debugg-ai-mcpcheck_app_in_browser
Runs an AI browser agent against your app. The agent navigates, interacts, and reports back with screenshots.
| Parameter | Type | Description |
|---|---|---|
description | string required | What to test (natural language) |
url | string | Target URL — required if localPort not set |
localPort | number | Local dev server port — tunnel created automatically |
environmentId | string | UUID of a specific environment |
credentialId | string | UUID of a specific credential |
credentialRole | string | Pick a credential by role (e.g. admin, guest) |
username | string | Username for login |
password | string | Password for login |
Configuration
bash
DEBUGGAI_API_KEY=your_api_keyLocal Development
bash
npm install && npm test && npm run buildLinks
Dashboard · Docs · Issues · Discord
---
Apache-2.0 License © 2025 DebuggAI
Similar MCP
Based on tags & features
Trending MCP
Most active this week