Patchright LiteSAFE
Patchright (Playwright patch) MCP server for lightweight models
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
A streamlined Model Context Protocol (MCP) server that wraps the Patchright Node.js SDK to provide stealth browser automation capabilities to AI models. This lightweight server focuses on essential functionality to make it easier for simpler AI models to use.
What is Patchright?
Patchright is an undetected version of the Playwright testing and automation framework. It's designed as a drop-in replacement for Playwright, but with advanced stealth capabilities to avoid detection by anti-bot systems. Patchright patches various detection techniques including:
- Runtime.enable leak
- Console.enable leak
- Command flags leaks
- General detection points
- Closed Shadow Root interactions
This MCP server wraps the Node.js version of Patchright to make its capabilities available to AI models through a simple, standardized protocol.
Features
- Simple Interface: Focused on core functionality with just 4 essential tools
- Stealth Automation: Uses Patchright's stealth mode to avoid detection
- MCP Standard: Implements the Model Context Protocol for easy AI integration
- Stdio Transport: Uses standard input/output for seamless integration
Prerequisites
- Node.js 18+
- npm or yarn
Installation
- Clone this repository:
git clone https://github.com/yourusername/patchright-lite-mcp-server.git cd patchright-lite-mcp-server
- Install dependencies:
npm install
- Build the TypeScript code:
npm run build
- Install Chromium-Driver for Pathright:
npx patchright install chromium
Usage
Run the server with:
npm start
This will start the server with stdio transport, making it ready to integrate with AI tools that support MCP.
Integrating with AI Models
Claude Desktop
Add this to your claude-desktop-config.json file:
{
"mcpServers": {
"patchright": {
"command": "node",
"args": ["p59d00095b792OBSERVED · 2026-10-10Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.
claude mcp add patchright-lite-mcp-server -- npx -y [email protected]
{
"mcpServers": {
"patchright-lite-mcp-server": {
"command": "npx",
"args": [
"-y",
"[email protected]"
]
}
}
}Exposed tools (4)
4 read · 0 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
browse | read | Browse to a URL and return the page title and visible text |
close | read | Close browser to free resources |
extract | read | Extract information from the current page as text, html, or screenshot |
interact | read | Perform simple interactions on a page |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (1)
@modelcontextprotocol/sdk, patchright, uuid, zod, @types/node, typescript
Gates applied: no_behavioural_pass.
59d00095b792full audit observations/trust-audit/mcp-server/dylangroos__patchright-lite.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-10 | 59d00095b792 | SAFE | B | 89 | first audit |
Questions
What is the Patchright Lite MCP server?
Patchright (Playwright patch) MCP server for lightweight models
What tools does Patchright Lite expose?
4 in total: 4 read-only, 0 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is Patchright Lite safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does Patchright Lite need?
No credential environment variables were found in its source, so it appears to need none.
How does Patchright Lite run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as patchright-lite-mcp-server at 1.0.0.
How current is this page?
The grade is for one exact copy of the source (59d00095b792), read on 2026-10-10. The repository is watched and re-audited when it changes.