UniAuto MCP Server
About
Universal Test Automation MCP Server with self-healing capabilities and Smithery.ai integration
Details
- Author
- samuelvinay91
- GitHub stars
- 4
- Downloads
- 161
- Categories
- Other
Jump to
- Multi-layered self-healing automation with 5-tier recovery strategies
- AI-driven test generation via natural language and Claude integration
- MCP compatibility for direct AI assistant control
- Unified web and desktop automation with Playwright
- Visual element matching and image recognition
- Advanced testing: visual, accessibility, performance, network tracing
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
UniAuto MCP ServerCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Clone the repository, install dependencies, install Playwright browsers, configure the .env file with a Claude API key, and start the server. Use Smithery.ai or direct MCP registration to connect with AI assistants like Claude.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"uniauto mcp server": {
"uniauto-mcp-server": {
"command": "npx",
"args": [
"playwright",
"install"
]
}
}
}
}
McpServers
{
"uniauto-mcp-server": {
"command": "npx",
"args": [
"playwright",
"install"
]
}
}
UniAuto MCP Server
<p align="center">

</p>
Universal Test Automation MCP Server for web and desktop automation with self-healing capabilities and AI integration.
Cutting-Edge Features
UniAuto MCP Server incorporates several cutting-edge and futuristic features that position it at the forefront of test automation technology:
1. Multi-Layered Self-Healing Automation
The server's 5-tier self-healing mechanism is revolutionary, using a sophisticated cascade of recovery strategies: - Repository-based alternative selectors - Role-based accessibility selectors - Dynamic looser selector generation - Visual element matching using DOM and image recognition - Contextual text-proximity detectionThis multi-strategy approach creates extraordinarily resilient tests that can survive even major UI overhauls.
2. AI-Driven Test Generation & Maintenance
Integration with Claude and other LLMs enables: - Natural language test creation ("Automate the checkout process") - Automatic test step generation with appropriate selectors - Intelligent response parsing to turn AI outputs into executable commands - Visual context interpretation (using screenshots as inputs) - Automated test code generation in multiple frameworks (Playwright, Cypress, Jest, etc.) - Support for various testing styles (BDD, TDD) and output formats - Application structure analysis for intelligent test creation3. MCP (Model Context Protocol) Integration
The MCP implementation is particularly forward-thinking: - Allows AI assistants to directly control testing infrastructure - Exposes a standardized API for automation actions - Enables seamless integration with Smithery.ai and other AI orchestration platforms - Creates a bridge between LLMs and web/desktop automation4. Unified Web + Desktop Automation
The integration of both web testing (via Playwright) and desktop automation within a single framework provides: - Cross-domain testing capabilities - Ability to test flows that span browsers and native applications - Consistent API across different application types5. Visual Intelligence
The implementation of visual matching provides: - Element recognition by appearance rather than structure - Testing continuity even when DOM elements change completely - Pixel-perfect comparison capabilitiesCore Features
- Web automation using Playwright (multi-browser support)
- Desktop automation capabilities
- Self-healing selectors for resilient test automation
- AI integration for test generation (supports Claude and other models)
- Automated test code generation in multiple frameworks and styles
- MCP (Model Context Protocol) compatibility for use with AI assistants
- Test case management and execution history
- Visual element matching
Impact on Test Automation Ecosystem
UniAuto MCP Server has the potential to transform the test automation landscape:
1. Maintenance Cost Reduction
The self-healing capabilities dramatically reduce the #1 pain point in test automation - maintenance costs. Studies suggest that organizations spend 30-40% of testing time on test maintenance. UniAuto could reduce this by 70-80%, freeing QA resources for more strategic work.2. Democratization of Test Automation
The AI integration allows non-technical stakeholders to create tests using natural language, potentially expanding test automation adoption by: - Allowing product managers to directly specify acceptance tests - Enabling business analysts to validate features without coding - Reducing the specialized skills required for test creation3. Resilience Revolution
By solving the brittleness problem in modern automation, UniAuto shifts the entire ecosystem toward resilience-focused testing rather than selector-focused testing. This paradigm shift fundamentally changes how tests are written and evaluated.4. AI Amplification in Testing
As an MCP-compatible system that works with Claude and other models, UniAuto positions automation as an "AI-native" technology, creating a foundation for: - AI-driven exploratory testing - Autonomous test maintenance - Intelligent test prioritization based on application changes - Natural language test reporting for stakeholders5. Cross-Domain Testing Standards
By unifying web and desktop testing under one framework, UniAuto helps establish new standards for end-to-end testing that transcend the current siloed approach to different application types.Installation
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.



