Skip to main content
Glama
vincenthopf

Gemini Web Automation MCP

by vincenthopf
OverviewSchemaRelated ServersScoreDiscussions
Python
Local

Gemini Web Automation MCP

PyPI Python Tests License: MIT MCP

Production-ready Model Context Protocol (MCP) server providing AI-powered web browsing automation using Google's Gemini 2.5 Computer Use API. Built with FastMCP and optimized for 4-5x faster performance than baseline implementations.

Table of Contents

Related MCP server: Cloudflare Playwright MCP

Overview

Gemini Web Automation MCP enables Claude Desktop and other MCP clients to perform intelligent web browsing automation. The AI agent navigates websites, clicks buttons, fills forms, searches for information, and interacts with web pages like a human user.

Key Statistics:

  • 7 MCP tools for comprehensive browser control

  • 13 browser actions supported (click, type, scroll, navigate, etc.)

  • 4-5x faster than naive implementations

  • 90% context reduction with compact progress mode

  • 1440x900 optimized resolution (Gemini recommended)

Features

Core Capabilities

7 Production-Ready MCP Tools:

  • browse_web - Synchronous web browsing with immediate completion

  • start_web_task - Start long-running tasks in background

  • check_web_task - Monitor progress with compact/full modes

  • wait - Intelligent rate limiting (1-60 seconds)

  • stop_web_task - Cancel running background tasks

  • list_web_tasks - View all active and completed tasks

  • get_web_screenshots - Retrieve session screenshots for verification

Advanced Features:

  • Real-time progress tracking with timestamped events

  • Automatic screenshot capture at each step

  • Safety decision framework (Gemini safety controls)

  • Context-aware polling with recommended delays

  • Background task management with status tracking

  • Session-based screenshot organization

Technical Highlights

  • MCP Protocol Compliant: Follows 2025 Model Context Protocol best practices

  • Performance Optimized: Conditional wait states (0.3-3s), fast page loads, single screenshot per turn

  • Safety-First: Implements Gemini's safety decision acknowledgment framework

  • Production Ready: Comprehensive error handling, logging, and validation

  • User-Friendly: Action-oriented tool names, clear descriptions, helpful examples

Quick Start

Prerequisites

Installation (Super Simple!)

Option A: One-Line Install (PyPI)

// Add to Claude Desktop config
{
  "mcpServers": {
    "computer-use": {
      "command": "uvx",
      "args": ["computer-use-mcp"],
      "env": {"GEMINI_API_KEY": "your_key_here"}
    }
  }
}

Then: playwright install chromium and restart Claude Desktop. Done!

Option B: Local Development

git clone https://github.com/vincenthopf/computer-use-mcp.git
cd computer-use-mcp
uv sync
playwright install chromium
cp .env.sample .env
# Edit .env with your GEMINI_API_KEY
uv run mcp dev server.py

The MCP Inspector will open at http://localhost:6274 where you can test all tools interactively.

Verification

Run validation tests to ensure everything is set up correctly:

uv run python3 test_server.py

All tests should pass with ✓ markers.

Installation

Method 1: UVX (Recommended - One Command)

The simplest way to use this MCP server. Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "computer-use": {
      "command": "uvx",
      "args": ["computer-use-mcp"],
      "env": {
        "GEMINI_API_KEY": "your_api_key_here"
      }
    }
  }
}

That's it! uvx automatically downloads and installs the package from PyPI. Restart Claude Desktop and you're ready to go.

Note: You'll also need to install Playwright browsers once:

playwright install chromium

Method 2: Local Development

For development or contributing:

git clone https://github.com/vincenthopf/computer-use-mcp.git
cd computer-use-mcp
uv sync
playwright install chromium
cp .env.sample .env
# Edit .env with your GEMINI_API_KEY
uv run mcp dev server.py

Method 3: Direct Git Install

Install directly from GitHub without cloning:

{
  "mcpServers": {
    "computer-use": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/vincenthopf/computer-use-mcp.git", "computer-use-mcp"],
      "env": {
        "GEMINI_API_KEY": "your_api_key_here"
      }
    }
  }
}

Configuration

Create a .env file in the project root with the following variables:

Required Configuration

# Gemini API Configuration
GEMINI_API_KEY=your_api_key_here
GEMINI_MODEL=gemini-2.5-computer-use-preview-10-2025

Optional Configuration

# Browser Configuration
SCREEN_WIDTH=1440        # Recommended by Google (don't change)
SCREEN_HEIGHT=900        # Recommended by Google (don't change)
HEADLESS=false           # Set to 'true' for faster headless mode

# Output Configuration (optional - defaults to system temp directory)
# SCREENSHOT_OUTPUT_DIR=/custom/path/to/screenshots

Note: Screen resolution of 1440x900 is optimized for Gemini's Computer Use model. Other resolutions may degrade performance.

Usage

Synchronous Workflow

For quick tasks that complete in under 10 seconds, use the synchronous browse_web tool:

# Example: Quick product search
result = browse_web(
    task="Go to Amazon and find the top 3 gaming laptops with prices",
    url="https://www.amazon.com"
)

# Response includes full results and progress history
{
    "ok": true,
    "data": "Found top 3 gaming laptops...",
    "session_id": "20251017_143022_abc123",
    "screenshot_dir": "output_screenshots/20251017_143022_abc123",
    "progress": [
        {"timestamp": "...", "type": "info", "message": "Started browser automation"},
        {"timestamp": "...", "type": "turn", "message": "Turn 1/30"},
        {"timestamp": "...", "type": "function_call", "message": "Action: navigate"}
    ]
}

Asynchronous Workflow

For long-running tasks (15+ seconds), use the async workflow to monitor progress:

Step 1: Start the task

result = start_web_task(
    task="Research top 10 AI companies and compile details",
    url="https://www.google.com"
)
# Returns immediately with task_id
# {"task_id": "abc-123-def", "status": "running"}

Step 2: Check progress

# Wait 5 seconds
wait(5)

# Check status (compact mode - recommended)
status = check_web_task(task_id="abc-123-def", compact=true)

# Response shows recent progress summary
{
    "ok": true,
    "task_id": "abc-123-def",
    "status": "running",
    "progress_summary": {
        "total_steps": 12,
        "recent_actions": [
            "Turn 8/30",
            "Action: type_text_at",
            "Action: click_at"
        ]
    },
    "recommended_poll_after": "2025-01-17T14:30:15Z"
}

Step 3: Get final result

# Continue polling until status is "completed"
final = check_web_task(task_id="abc-123-def")

# When completed, result field contains data
{
    "ok": true,
    "status": "completed",
    "result": {
        "ok": true,
        "data": "Compiled research on 10 AI companies...",
        "session_id": "..."
    }
}

All 7 MCP Tools

1. browse_web

Synchronous web browsing that waits for task completion.

Parameters:

  • task (string, required): Natural language description of what to accomplish

  • url (string, optional): Starting URL (defaults to Google)

Returns:

  • ok (boolean): Success status

  • data (string): Task completion message with results

  • session_id (string): Unique session identifier

  • screenshot_dir (string): Path to saved screenshots

  • progress (array): Full progress history with timestamps

  • error (string): Error message if task failed

Example:

browse_web(
    task="Go to example.com and click the login button",
    url="https://example.com"
)

2. start_web_task

Start a long-running web task in the background.

Parameters:

  • task (string, required): Natural language task description

  • url (string, optional): Starting URL

Returns:

  • ok (boolean): Task started successfully

  • task_id (string): Unique ID for checking progress

  • status (string): Always "running" initially

  • message (string): Instructions for checking progress

Example:

start_web_task(
    task="Research and compare prices for iPhone 15 on 5 different sites"
)

3. check_web_task

Monitor progress of a background task.

Parameters:

  • task_id (string, required): Task ID from start_web_task

  • compact (boolean, optional): Return summary only (default: true)

Returns (compact mode):

  • ok (boolean): Success status

  • task_id (string): Task identifier

  • status (string): "pending" | "running" | "completed" | "failed" | "cancelled"

  • progress_summary (object): Recent actions and total steps

  • result (object): Task results (when completed)

  • error (string): Error message (when failed)

  • recommended_poll_after (string): ISO timestamp for next check

Example:

check_web_task(task_id="abc-123", compact=true)

4. wait

Pause execution for rate limiting and polling delays.

Parameters:

  • seconds (integer, required): Wait time in seconds (1-60)

Returns:

  • ok (boolean): Success status

  • waited_seconds (integer): Actual wait time

  • message (string): Confirmation message

Example:

wait(5)  # Wait 5 seconds between status checks

5. stop_web_task

Cancel a running background task.

Parameters:

  • task_id (string, required): Task ID to cancel

Returns:

  • ok (boolean): Cancellation success

  • message (string): Confirmation message

  • task_id (string): Cancelled task ID

  • error (string): Error if task not found

Example:

stop_web_task(task_id="abc-123")

6. list_web_tasks

View all tasks (active and completed).

Parameters: None

Returns:

  • ok (boolean): Success status

  • tasks (array): List of task status objects

  • count (integer): Total number of tasks

  • active_count (integer): Number of running tasks

Example:

list_web_tasks()

7. get_web_screenshots

Retrieve screenshots from a completed session.

Parameters:

  • session_id (string, required): Session ID from browse_web or check_web_task

Returns:

  • ok (boolean): Success status

  • screenshots (array): List of screenshot file paths

  • session_id (string): Session identifier

  • count (integer): Number of screenshots

  • error (string): Error if session not found

Example:

get_web_screenshots(session_id="20251017_143022_abc123")

Architecture

┌─────────────────────────────────────────┐
│    MCP Client (Claude Desktop, etc.)    │
└─────────────────┬───────────────────────┘
                  │ MCP Protocol (JSON-RPC)
┌─────────────────▼───────────────────────┐
│           FastMCP Server                │
│  7 Tools: browse_web, start_web_task,  │
│  check_web_task, wait, stop_web_task,  │
│  list_web_tasks, get_web_screenshots   │
└─────────────────┬───────────────────────┘
                  │
┌─────────────────▼───────────────────────┐
│       BrowserTaskManager                │
│  Manages async task queue, status      │
│  tracking, and background threads       │
└─────────────────┬───────────────────────┘
                  │
┌─────────────────▼───────────────────────┐
│      GeminiBrowserAgent                 │
│  Executes browser automation loop:     │
│  Screenshot → Gemini → Actions → Loop  │
└─────────────────┬───────────────────────┘
                  │
        ┌─────────┴─────────┐
        │                   │
┌───────▼────────┐  ┌──────▼────────┐
│  Gemini API    │  │  Playwright   │
│  (Vision+AI)   │  │  (Chromium)   │
└────────────────┘  └───────────────┘

How It Works

  1. Task Submission: MCP client sends natural language task + optional URL

  2. Browser Launch: Playwright launches Chromium (1440x900 viewport)

  3. Gemini Loop: Screenshot → Gemini vision analysis → Browser actions → Screenshot (repeat)

  4. Completion: Gemini returns text response when task is complete

  5. Cleanup: Browser closed, screenshots saved to session directory

Browser Actions

The Gemini Computer Use API supports 13 browser automation actions:

Action

Description

Example Use

navigate

Go to a URL

Navigate to https://example.com

click_at

Click at normalized coordinates (0-999)

Click button at position (500, 300)

hover_at

Hover at normalized coordinates

Hover over menu item

type_text_at

Type text at coordinates

Type search query in input field

key_combination

Press keyboard combinations

Press Enter, Ctrl+A, etc.

scroll_document

Page-level scrolling

Scroll down one page

scroll_at

Scroll at specific coordinates

Scroll within specific element

drag_and_drop

Drag from source to destination

Reorder list items

go_back

Navigate backward in history

Return to previous page

go_forward

Navigate forward in history

Go to next page

search

Navigate to Google search

Start a Google search

wait_5_seconds

Wait 5 seconds

Wait for dynamic content

open_web_browser

No-op (browser already open)

-

Coordinate System: Gemini uses a normalized 1000x1000 grid. The agent automatically converts to actual pixels based on your screen resolution.

Performance

Benchmark Results

Task Type

Duration

Turns

Description

Simple

5-8 seconds

1-2

Navigate to a page, verify content

Medium

20-40 seconds

5-8

Search, click, extract information

Complex

60-120 seconds

15-30

Multi-step workflows, data compilation

Optimization Strategies

Before Optimization:

  • 6 seconds wait after every action

  • networkidle page loads (waits for all network requests)

  • 3 screenshots captured per turn

  • Sequential action execution

After Optimization:

  • 0.3-3 seconds conditional waits (navigation only)

  • domcontentloaded page loads (DOM ready)

  • 1 screenshot per turn (reused for all responses)

  • Parallel function execution when possible

Performance Improvement: 4-5x faster execution

Latency Breakdown (Per Turn)

  • Gemini API call: 2-4 seconds (vision processing)

  • Browser action: 0.3-3 seconds (optimized waits)

  • Screenshot capture: <0.5 seconds

  • Total per turn: ~3-8 seconds

Best Practices

Task Design

1. Be Specific and Clear

# ❌ Bad: Vague task
"Search for stuff"

# ✅ Good: Specific task
"Go to Amazon, search for 'wireless headphones', and find the top 3 results with prices"

2. Choose the Right Tool

# ❌ Bad: Async for simple tasks
start_web_task("Navigate to google.com")  # Requires 3+ tool calls

# ✅ Good: Sync for simple tasks
browse_web("Navigate to google.com")  # Single tool call

3. Use Appropriate Polling

# ❌ Bad: Poll too frequently
check_web_task(task_id)  # Called every 0.5 seconds

# ✅ Good: Poll every 3-5 seconds
wait(5)
check_web_task(task_id)

Error Handling

4. Handle All Status States

status = check_web_task(task_id)

if status["status"] == "completed":
    process_result(status["result"])
elif status["status"] == "failed":
    handle_error(status["error"])
elif status["status"] == "running":
    continue_polling()

5. Use Compact Mode

# ✅ Recommended: Compact mode (90% smaller)
check_web_task(task_id, compact=true)

# ⚠️ Only when needed: Full progress
check_web_task(task_id, compact=false)

Security

6. Review Screenshots Always check saved screenshots to verify agent behavior, especially for sensitive operations.

7. Environment Variables Never commit API keys. Always use environment variables or secure vaults.

8. Rate Limiting Use the wait tool to respect rate limits and avoid overwhelming services.

9. Domain Validation Be cautious with user-provided URLs. Consider implementing domain allowlists for production.

10. Logging and Audit All actions are logged with timestamps. Review logs for debugging and compliance.

Troubleshooting

Common Issues

Error: 400 INVALID_ARGUMENT (safety decision)

  • Solution: This is fixed in v1.0.0. Update to latest version.

Context window filling up too fast

  • Solution: Use compact=true in check_web_task (default behavior).

Tasks timing out at 30 turns

  • Solution: Break complex tasks into smaller subtasks or increase max_turns in browser_agent.py.

Browser not visible during execution

  • Solution: Set HEADLESS=false in .env to see the browser window.

"No module named 'mcp'" error

  • Solution: Activate virtual environment and run uv sync.

"GEMINI_API_KEY environment variable not set"

  • Solution: Create .env file with your API key or set it in Claude Desktop config.

"Executable doesn't exist" (Playwright)

  • Solution: Run playwright install chromium.

MCP server not showing in Claude Desktop

  • Solution: Verify absolute paths in config, ensure uv is in PATH, restart Claude Desktop completely.

FAQ

Q: Can I use both synchronous and asynchronous workflows? A: Yes! Use browse_web for quick tasks (<10s) and start_web_task for long tasks (>15s).

Q: What happens to old completed tasks? A: Tasks auto-cleanup after 24 hours to free memory.

Q: Can I check progress of a synchronous task? A: No, but the response includes full progress history after completion.

Q: How many tasks can run simultaneously? A: Unlimited. Each task runs in its own browser instance and thread.

Q: Does Claude automatically poll async tasks? A: No, Claude must manually call check_web_task() multiple times with wait() between calls.

Q: Can I cancel a task mid-execution? A: Yes, use stop_web_task(task_id) to cancel any running task.

Development

Setup for Contributors

1. Fork and clone:

git clone https://github.com/YOUR_USERNAME/gemini-web-automation-mcp.git
cd gemini-web-automation-mcp

2. Install dependencies:

uv sync
playwright install chromium

3. Create .env:

cp .env.sample .env
# Add your GEMINI_API_KEY

4. Test your setup:

uv run mcp dev server.py

Testing

Manual Testing:

# Start MCP Inspector
uv run mcp dev server.py

# Opens web interface at http://localhost:6274
# Test all tools interactively

Validation Tests:

# Run automated validation
uv run python3 test_server.py

Coding Standards

  • Follow PEP 8 style guide

  • Use type hints for all function signatures

  • Write docstrings for public functions/classes

  • Keep functions focused and small

  • Use meaningful variable names

Example Function:

async def check_web_task(task_id: str, compact: bool = True) -> dict[str, Any]:
    """
    Check progress of a background web browsing task.

    Args:
        task_id: Task ID from start_web_task()
        compact: Return summary only (default: True)

    Returns:
        Dictionary containing task status and progress

    Raises:
        ValueError: If task_id is invalid
    """
    # Implementation
    pass

Contributing

How to Contribute:

  1. Fork the repository

  2. Create a feature branch: git checkout -b feature/amazing-feature

  3. Make your changes following coding standards

  4. Test your changes with MCP Inspector

  5. Update documentation if needed

  6. Commit with clear message: git commit -m "Add: Brief description"

  7. Push to your fork: git push origin feature/amazing-feature

  8. Open a Pull Request

Commit Message Prefixes:

  • Add: - New features

  • Fix: - Bug fixes

  • Update: - Updates to existing features

  • Refactor: - Code refactoring

  • Docs: - Documentation changes

  • Test: - Adding or updating tests

Pull Request Checklist:

  • All tests pass

  • Documentation updated

  • Code follows style guidelines

  • No breaking changes (or clearly documented)

  • Commit messages are clear

Tool Design Principles

When adding or modifying MCP tools:

  1. User-focused naming: browse_web not execute_browser_automation

  2. Clear descriptions: Explain what users accomplish, not technical details

  3. Action-oriented: Use verbs (check, start, stop, wait)

  4. Proper validation: Validate inputs and provide helpful error messages

  5. Consistent responses: Always return {"ok": bool, ...} format

Deployment

Pre-Deployment Checklist

Project Files:

  • README.md with comprehensive documentation

  • LICENSE (MIT)

  • CHANGELOG.md

  • pyproject.toml with proper metadata

  • .gitignore with comprehensive rules

  • .env.sample template

  • Core files (server.py, browser_agent.py, task_manager.py)

Code Quality:

  • All tests passing

  • Safety decision bug fixed

  • Compact progress mode implemented

  • MCP best practices followed

  • Performance optimizations applied

GitHub Release Process

1. Create GitHub Repository:

# On GitHub, create new repository "gemini-web-automation-mcp"
# Do NOT initialize with README (we have it)

2. Push to GitHub:

git remote add origin https://github.com/YOUR_USERNAME/gemini-web-automation-mcp.git
git branch -M main
git push -u origin main

3. Create Release Tag:

git tag -a v1.0.0 -m "Version 1.0.0 - Initial production release"
git push origin v1.0.0

4. Create GitHub Release:

  • Go to Releases → Create a new release

  • Choose tag: v1.0.0

  • Release title: v1.0.0 - Initial Production Release

  • Description: Copy from CHANGELOG.md

  • Publish release

Repository Settings

Configure:

  • Description: Production-ready MCP server for AI-powered web automation

  • Topics: mcp, gemini, browser-automation, claude-desktop, ai-agents, playwright

  • Enable Issues for bug reports

  • Enable Discussions for community support

Distribution Methods

Method 1: Direct Git Clone (Recommended)

git clone https://github.com/yourusername/gemini-web-automation-mcp.git
cd gemini-web-automation-mcp
uv sync
playwright install chromium

Method 2: UVX (Future - when published to PyPI)

uvx gemini-web-automation-mcp

Roadmap

Planned Features

Priority 1: Security Enhancements

  • Human-in-the-loop confirmation UI for safety decisions

  • Domain allowlist/blocklist for navigation

  • Input sanitization and validation

  • Container-based sandboxing for production

Priority 2: Reliability

  • Retry logic with exponential backoff

  • Better error recovery mechanisms

  • Network resilience improvements

  • Connection timeout handling

Priority 3: Functionality

  • Cookie and session management

  • Form auto-fill templates

  • Multi-tab support

  • Mobile viewport emulation

Priority 4: Developer Experience

  • Proxy support for corporate environments

  • Custom wait conditions

  • Advanced screenshot comparison

  • Performance profiling tools

Known Limitations

  • Maximum 30 turns per task (configurable but not recommended to increase)

  • Browser automation only (no desktop OS-level control)

  • Single browser instance per task (no tab switching)

  • Limited to Chromium (Firefox/WebKit not supported)

  • Safety confirmations not yet implemented (future enhancement)

Changelog

[1.0.0] - 2025-01-17

Initial production-ready release

Added:

  • 7 MCP tools for comprehensive browser automation

  • Real-time progress tracking with compact mode (90% size reduction)

  • Safety decision framework (fixes 400 INVALID_ARGUMENT error)

  • Context-aware polling with recommended delay timestamps

  • Performance optimizations (4-5x faster than baseline)

  • Automatic screenshot capture at each step

  • Background task management with status tracking

  • Comprehensive documentation and examples

Performance:

  • Conditional wait states (0.3-3s vs 6s per action)

  • Fast page loads (domcontentloaded vs networkidle)

  • Single screenshot per turn (eliminates duplicates)

  • Parallel function execution for batch operations

Security:

  • Safety decision acknowledgment implementation

  • Environment-based API key configuration

  • Comprehensive logging for audit trails

  • Screenshot-based verification capability

Full changelog →

Security

Safety Decision Framework

This MCP server implements Gemini's safety decision framework to prevent:

  • Financial transactions without confirmation

  • Sensitive data access without review

  • System-level changes without approval

  • CAPTCHA bypassing attempts

  • Potentially harmful actions

Best Practices

  1. Sandboxing: Run in containerized environment for production deployment

  2. API Key Security: Use environment variables, never commit keys to version control

  3. Rate Limiting: Respect built-in polling delays (recommended 5 seconds)

  4. Human-in-Loop: Review outputs before taking actions based on results

  5. Logging: All actions are logged with timestamps for audit trail

  6. Screenshot Review: Check saved screenshots to verify agent behavior

Reporting Vulnerabilities

For security vulnerabilities, please email security@example.com with:

  • Description of the vulnerability

  • Steps to reproduce

  • Potential impact

  • Suggested fix (if available)

Do not open public issues for security vulnerabilities.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

Built with excellence using:

  • Google Gemini Team - Gemini 2.5 Computer Use API

  • Anthropic - Model Context Protocol specification

  • FastMCP - Excellent MCP server framework

  • Playwright - Robust browser automation library

Support

Get Help

Useful Links

Community

  • Share your use cases and automations

  • Contribute improvements and bug fixes

  • Help others in GitHub Discussions

  • Star the repository if you find it useful

Built with Gemini 2.5 Computer Use API

MCP Protocol 2025 | Production Ready | Open Source

Available Tools

7 tools
browse_webA
Browse the web to complete a task using AI-powered browser automation.

The AI agent can navigate websites, click buttons, fill forms, search for information,
and interact with web pages just like a human user. This runs synchronously and returns
when the task is complete.

Args:
    task: What you want to accomplish (e.g., "Find the top 3 gaming laptops on Amazon")
    url: Starting webpage (defaults to Google)

Returns:
    Dictionary containing:
    - ok: Boolean indicating success
    - data: Task completion message with results
    - screenshot_dir: Path to saved screenshots
    - session_id: Unique session identifier
    - progress: List of actions taken during browsing
    - error: Error message (if task failed)

Examples:
    - "Search for Python tutorials and summarize the top result"
    - "Go to example.com and click the login button"
    - "Find product reviews for iPhone 15 Pro"

Note: For long-running tasks, consider using start_web_task instead.
ParametersJSON Schema
NameRequiredDescriptionDefault
taskYes
urlNohttps://www.google.com

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.8/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It clearly describes that the tool runs synchronously, returns when the task is complete, and can perform human-like interactions. However, it doesn't mention rate limits, authentication requirements, or potential destructive actions beyond what's implied by 'interact with web pages.'

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured with clear sections (purpose, behavior, args, returns, examples, note). Every sentence adds value, and it's appropriately sized for a complex tool with multiple parameters and return values.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (web automation with 2 parameters), no annotations, and the presence of an output schema, the description provides excellent context. It explains the tool's behavior, parameters, return structure, usage guidelines, and includes examples, making it complete enough for an agent to use effectively.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 0% schema description coverage, the description fully compensates by providing detailed parameter semantics. It explains the 'task' parameter with examples and clarifies that 'url' defaults to Google. This adds significant value beyond the bare schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose as 'browse the web to complete a task using AI-powered browser automation' with specific verbs (navigate, click, fill, search, interact) and distinguishes it from sibling tools by mentioning it runs synchronously and returns when complete. It explicitly differentiates from start_web_task for long-running tasks.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit guidance on when to use this tool (for synchronous web browsing tasks) and when not to use it (for long-running tasks, consider using start_web_task instead). It also provides three concrete examples of appropriate use cases.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

check_web_taskA
Check progress of a background web browsing task.

Returns a summary of task progress. By default, returns compact format to
avoid filling your context window with verbose progress logs.

IMPORTANT: To prevent context bloat, wait at least 3-5 seconds between
checks. Use the 'recommended_poll_after' timestamp as guidance.

Args:
    task_id: Task ID from start_web_task()
    compact: Return summary only (default: True). Set to False for full details.

Returns:
    Dictionary containing:
    - ok: Boolean indicating success
    - task_id: Task identifier
    - status: "pending", "running", "completed", "failed", or "cancelled"
    - progress_summary: Recent actions (compact mode only)
    - progress: Full action history (full mode only)
    - result: Task results (when completed)
    - error: Error message (when failed)
    - recommended_poll_after: Timestamp to check again (when running)
    - polling_guidance: Message about polling frequency

Examples:
    - check_web_task("abc-123-def")  # Compact summary
    - check_web_task("abc-123-def", compact=False)  # Full details

Best Practice:
    Only poll every 3-5 seconds to keep your context window clean.
    Use the wait() tool to pause between checks if your platform doesn't
    support automatic delays.

Recommended workflow:
    1. start_web_task("...")
    2. wait(5)
    3. check_web_task(task_id)
    4. If still running, repeat steps 2-3
ParametersJSON Schema
NameRequiredDescriptionDefault
task_idYes
compactNo

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.9/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure and does so comprehensively. It explains the tool's polling behavior (3-5 second intervals), context management strategy (compact format to avoid bloat), return format variations based on parameters, and provides practical guidance about using the wait() tool for delays.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured with clear sections (purpose, args, returns, examples, best practice, workflow) but could be slightly more concise. The 'Best Practice' and 'Recommended workflow' sections contain some redundancy about polling intervals. Every sentence adds value, but some information is repeated.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (polling behavior, output variations) and the presence of an output schema, the description provides excellent contextual completeness. It explains the tool's role in the workflow, behavioral constraints, parameter effects, and practical usage patterns while appropriately deferring detailed return structure to the output schema.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 0% schema description coverage, the description fully compensates by explaining both parameters in detail. It clarifies that task_id comes from start_web_task(), and explains the compact parameter's effect on output format (summary vs full details) with clear examples showing both usage patterns.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose with specific verbs ('check progress', 'returns a summary') and distinguishes it from siblings by explicitly mentioning it works with tasks created by 'start_web_task()'. It identifies the exact resource being operated on (background web browsing tasks).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit guidance on when to use this tool (after starting a task with start_web_task, with 3-5 second intervals between checks) and includes a complete recommended workflow. It also distinguishes this from other tools by showing its role in the task lifecycle.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

get_web_screenshotsA
Retrieve screenshots captured during a web browsing session.

Each browsing session saves screenshots of the pages visited. Use this to
review what the AI agent saw and did during task execution.

Args:
    session_id: Session ID returned from browse_web or check_web_task

Returns:
    Dictionary containing:
    - ok: Boolean indicating success
    - screenshots: List of screenshot file paths
    - session_id: The session identifier
    - count: Number of screenshots found
    - error: Error message (if session not found)

Example:
    get_web_screenshots("20251017_143022_a1b2c3d4")
ParametersJSON Schema
NameRequiredDescriptionDefault
session_idYes

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.2/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It discloses that the tool retrieves saved screenshots and returns a dictionary with specific fields, including error handling for 'session not found.' However, it lacks details on permissions, rate limits, or whether the operation is read-only or has side effects, which are important for a tool interacting with session data.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured with a clear purpose statement, usage context, parameter details, return values, and an example. It is appropriately sized, but the example could be integrated more seamlessly, and some sentences like 'Each browsing session saves screenshots...' slightly extend beyond strict necessity.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's moderate complexity, no annotations, and an output schema (implied by the Returns section), the description is mostly complete. It covers purpose, usage, parameters, and return values. However, it lacks details on behavioral aspects like error conditions beyond 'session not found' or performance considerations, leaving minor gaps.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 0%, so the description must compensate. It explicitly documents the single parameter: 'session_id: Session ID returned from browse_web or check_web_task,' adding crucial meaning by specifying the source of the session_id and its purpose, which the schema alone does not provide.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: 'Retrieve screenshots captured during a web browsing session.' It specifies the verb ('retrieve'), resource ('screenshots'), and context ('web browsing session'), distinguishing it from siblings like browse_web or check_web_task by focusing on post-session retrieval rather than active browsing or task management.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides clear context for when to use this tool: 'Each browsing session saves screenshots... Use this to review what the AI agent saw and did during task execution.' It implies usage after a session has ended, but does not explicitly state when not to use it or name alternatives, such as checking if a session exists first.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

list_web_tasksA
List all web browsing tasks, including active and completed ones.

Shows a summary of all tasks in the current session. Useful for tracking
multiple concurrent browsing operations.

Returns:
    Dictionary containing:
    - ok: Boolean indicating success
    - tasks: Array of task status objects (compact format)
    - count: Total number of tasks
    - active_count: Number of currently running tasks

Examples:
    - list_web_tasks()

Note: Returns compact task summaries. Use check_web_task(task_id) for details.
ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A3.8/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It describes the return format in detail, including specific keys like 'ok,' 'tasks,' 'count,' and 'active_count,' which adds value beyond basic listing. However, it doesn't mention potential limitations such as rate limits, session dependencies, or error handling, leaving some behavioral aspects unclear.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured and front-loaded with the core purpose. It uses bullet points for the return values and includes a note and example, which are helpful. However, some sentences could be more concise, such as 'Shows a summary of all tasks in the current session,' which slightly repeats the initial statement, but overall it's efficient and informative.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's low complexity (0 parameters) and the presence of an output schema (implied by the detailed return description), the description is quite complete. It explains the purpose, usage, return format, and provides an example and alternative tool. The only minor gap is the lack of explicit behavioral constraints like error cases, but this is mitigated by the output details.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema has 0 parameters with 100% coverage, so the schema fully documents the lack of inputs. The description adds no parameter information, which is appropriate here. Since there are no parameters, the baseline is 4, as the description doesn't need to compensate for any gaps and focuses on output semantics instead.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: 'List all web browsing tasks, including active and completed ones.' It specifies the verb ('List') and resource ('web browsing tasks'), and distinguishes it from siblings by mentioning 'compact task summaries' versus 'check_web_task(task_id) for details.' However, it doesn't explicitly differentiate from other list-like siblings if any existed, but since none are present, it's clear.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides clear context on when to use this tool: 'Shows a summary of all tasks in the current session. Useful for tracking multiple concurrent browsing operations.' It also gives an explicit alternative: 'Use check_web_task(task_id) for details.' However, it lacks explicit exclusions or comparisons to other siblings like 'browse_web' or 'start_web_task,' which could be helpful but isn't critical here.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

start_web_taskA
Start a web browsing task in the background and return immediately.

Use this for tasks that might take a while (30+ seconds). The task runs
asynchronously while you continue working. Check progress with check_web_task().

Args:
    task: What you want to accomplish on the web
    url: Starting webpage (defaults to Google)

Returns:
    Dictionary containing:
    - ok: Boolean indicating task was started successfully
    - task_id: Unique ID to check progress later
    - status: Will be "running"
    - message: Instructions for checking progress

Examples:
    - start_web_task("Research top 10 AI companies and their products")
    - start_web_task("Find and compare prices for MacBook Pro on 5 different sites")

Next steps:
    Use check_web_task(task_id) to monitor progress.
    Wait at least 5 seconds between status checks.
ParametersJSON Schema
NameRequiredDescriptionDefault
taskYes
urlNohttps://www.google.com

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.7/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes key traits: the task runs asynchronously in the background, returns immediately, requires monitoring with check_web_task, and has a default URL. However, it doesn't mention potential errors, timeouts, or resource limits, leaving some behavioral aspects uncovered.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured and appropriately sized, with clear sections (purpose, args, returns, examples, next steps). Every sentence adds value, such as explaining asynchronous behavior, providing usage examples, and outlining follow-up steps, with no redundant or wasted content.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (asynchronous operation with 2 parameters) and the presence of an output schema (which covers return values), the description is complete. It explains the tool's purpose, usage, parameters, and next steps adequately, compensating for the lack of annotations and low schema coverage.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 0%, so the description must compensate. It adds meaningful context for both parameters: 'task' is described as 'What you want to accomplish on the web' with examples, and 'url' is clarified as the 'Starting webpage (defaults to Google)'. This goes beyond the basic schema types, though it could provide more detail on URL formatting or task constraints.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the specific action ('Start a web browsing task in the background') and resource ('web browsing task'), distinguishing it from siblings like browse_web (likely synchronous) and check_web_task (monitoring). It explicitly mentions returning immediately and running asynchronously, which differentiates its purpose from other tools.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit guidance on when to use this tool ('for tasks that might take a while (30+ seconds)'), when not to use it (implied for shorter tasks), and alternatives (check_web_task for monitoring progress). It also specifies prerequisites like waiting 5 seconds between checks, making usage context clear.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

stop_web_taskA
Stop a running web browsing task.

Immediately halts task execution and cleans up browser resources. Use this
when you need to cancel a long-running task that's no longer needed.

Args:
    task_id: Task ID from start_web_task()

Returns:
    Dictionary containing:
    - ok: Boolean indicating success
    - message: Confirmation message
    - task_id: The stopped task ID
    - error: Error message (if task not found or already completed)

Examples:
    - stop_web_task("abc-123-def")

Note: Cannot stop tasks that are already completed or failed.
ParametersJSON Schema
NameRequiredDescriptionDefault
task_idYes

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.7/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes key behaviors: immediate halting of execution, cleanup of browser resources, and constraints on stopping completed/failed tasks. However, it doesn't mention potential side effects like data loss or error handling details beyond the return structure.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured and front-loaded with the core purpose, followed by usage guidelines, parameter details, return values, examples, and constraints. Every sentence adds value without redundancy, making it efficient and easy to parse.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (a destructive operation with one parameter), no annotations, and the presence of an output schema (which covers return values), the description is complete. It explains purpose, usage, parameters, returns, examples, and constraints, leaving no significant gaps for an agent to understand and invoke the tool correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 0%, so the description must compensate. It provides clear semantics for the single parameter (task_id from start_web_task()) and includes an example. While it doesn't detail format constraints beyond the example, it adds meaningful context beyond the bare schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the specific action ('stop', 'halts', 'cancels') and resource ('web browsing task'), distinguishing it from siblings like check_web_task, list_web_tasks, or wait. It explicitly mentions the tool stops a 'running' task, which differentiates it from tools that might handle completed or failed tasks.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit guidance on when to use this tool ('when you need to cancel a long-running task that's no longer needed') and when not to use it ('Cannot stop tasks that are already completed or failed'). It also implies alternatives by referencing start_web_task and other siblings for different operations.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

waitA
Wait for a specified number of seconds before continuing.

Use this when you need to pause between operations, such as:
- Waiting between status checks to avoid rapid polling
- Giving a web task time to make progress
- Rate limiting your requests
- Waiting for external processes to complete

Args:
    seconds: Number of seconds to wait (1-60)

Returns:
    Dictionary containing:
    - ok: Boolean indicating success
    - waited_seconds: How long the wait lasted
    - message: Confirmation message

Examples:
    - wait(5)  # Wait 5 seconds
    - wait(10)  # Wait 10 seconds

Best Practice:
    Use this instead of immediately polling check_web_task multiple times.
    Recommended wait time between status checks: 3-5 seconds.

Note: Maximum wait time is 60 seconds to prevent timeout issues.
ParametersJSON Schema
NameRequiredDescriptionDefault
secondsYes

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A5/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It thoroughly explains the tool's behavior, including the wait duration range (1-60 seconds), the return structure (dictionary with ok, waited_seconds, message), and practical constraints like timeout prevention. This goes beyond what the input schema alone provides.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured and front-loaded, starting with the core purpose, followed by usage guidelines, parameter details, returns, examples, and best practices. Every section adds value without redundancy, and the length is appropriate for a tool with behavioral complexity and no annotations.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's behavioral complexity (timing, constraints), lack of annotations, and the presence of an output schema, the description is complete. It covers purpose, usage, parameters, returns (though output schema exists, it adds clarification), examples, and best practices, leaving no gaps for an AI agent to understand and invoke the tool correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema has 0% description coverage, so the description must fully compensate. It adds significant meaning by explaining the 'seconds' parameter as 'Number of seconds to wait (1-60)', providing the semantic context, valid range, and examples (wait(5), wait(10)), which are not present in the schema's minimal title and type definition.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose with a specific verb ('wait') and resource ('specified number of seconds'), distinguishing it from sibling tools like check_web_task or start_web_task. It explicitly defines the action of pausing execution, which is distinct from the web-related operations of its siblings.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit guidance on when to use this tool, listing specific scenarios (e.g., avoiding rapid polling, rate limiting, waiting for external processes) and naming an alternative (check_web_task) in the 'Best Practice' section. It also includes when-not-to-use guidance by specifying a maximum wait time to prevent timeouts.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A4.5/5.0
Disambiguation5/5

Each tool has a distinct and clear purpose with no overlap. browse_web and start_web_task handle synchronous vs. asynchronous browsing, check_web_task monitors progress, get_web_screenshots retrieves visual data, list_web_tasks provides an overview, stop_web_task cancels tasks, and wait manages timing. The descriptions explicitly differentiate their roles, eliminating confusion.

Naming Consistency5/5

All tools follow a consistent verb_noun naming pattern (e.g., browse_web, check_web_task, get_web_screenshots). The pattern is uniform across all seven tools, with clear and descriptive names that align with their functions, making them predictable and easy to understand.

Tool Count5/5

With 7 tools, the count is well-scoped for web automation. It covers core operations like starting, monitoring, and stopping tasks, plus utilities like screenshots and waiting, without being excessive. Each tool serves a specific role, ensuring a balanced and functional set for the domain.

Completeness5/5

The toolset provides complete coverage for web automation workflows. It includes task initiation (browse_web, start_web_task), monitoring (check_web_task, list_web_tasks), management (stop_web_task), data retrieval (get_web_screenshots), and timing control (wait). There are no obvious gaps, supporting both synchronous and asynchronous operations effectively.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

  • MCP Browser Use Server

    JovaniPink
    A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI agents to interact with web browsers using natural language, featuring automated browsing, form filling, vision-based element detection, and structured JSON responses for systematic browser control.
    62
    MIT

  • Cloudflare Playwright MCP

    AI24K
    F
    license
    Not graded
    quality
    C
    maintenance
    Enables AI assistants to control a browser through a set of tools, allowing them to perform web automation tasks like navigation, typing, clicking, and taking screenshots.

  • Cloudflare Playwright MCP

    q275212
    F
    license
    Not graded
    quality
    C
    maintenance
    Enables AI assistants to control a browser through a set of tools, allowing them to perform web automation tasks like navigation, typing, clicking, and taking screenshots.

Appeared in Searches

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/vincenthopf/computer-use-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server