"Best open-source NPX or UVX tools for terminal deployment on Windows" matching MCP servers:

Enables MCP-capable agents to control a real QEMU virtual machine through nine computer-use tools—screenshot, click, move, type, key, scroll, status, start, and stop—without accessing the host’s pointer, clipboard, or files.

An MCP server that enables AI clients to autonomously discover, analyze, fix, test, submit, and track open-source contributions on GitHub through specialized tools.

Enables agent-driven CAD automation by inspecting, editing, and validating DWG drawings through a handle-first workflow, with support for visual grounding, CADPlan execution, and local SQLite workspace management.

Provides AI agents with tools for paper trading stocks, options, ETFs, and bonds, including advanced options strategies and risk analysis, using real market data without financial risk.

Enables AI coding agents to control real interactive shells, REPLs, and TUIs over stdio, supervise other agents, surface approval prompts, and verify outcomes without auto-approving or mistaking timeouts for success.

Exposes a Windows GUI agent as MCP tools so any MCP client can drive the real desktop through natural language: clicking, typing, scrolling, launching apps and running PowerShell, with fallback LLM providers for resilience. Also offers read-only inspection of open windows, the UI Automation tree and multi-monitor screenshots without invoking an LLM.

Delegate a Cursor task to a real, unattended background OpenCode agent — cheaper model, isolated history — through five small MCP tools.

Enables reproducible evaluation of AI agents on multi-source evidence investigation tasks, providing Docker-based scenarios, MCP tools, and automated scoring for payout reconciliation, performance-claim verification, and incident reconstruction.

A Windows-only MCP server that gives AI models eyes and hands on a Windows machine, enabling file reading, screenshots, mouse/keyboard injection, and an auto-debug loop with a safety-focused watchdog for protected abort regions.

Bridges the ChatGPT web UI to a local development environment over MCP, letting the model read and patch files, run real shell commands and terminals, execute tests, and control the desktop (screen, mouse, keyboard, clipboard) within user-approved folders. It adds parallel workers, goal/loop long-task control, plugin tools, and a security-hardening layer with risk-tiered shell policy, audit logging, and credential sanitization.

Enables AI hosts to preserve and resume task and project state across conversations, process restarts, and worker or model replacement by storing structured records, snapshots, receipts, and bounded context views through a local stdio interface.

Enables Claude Code, Cursor, and Codex users to orchestrate plan/build/review workflows across host or CLI models with a spec gate, tool vetting, hard per-run and per-day cost/token limits, safe defaults, and clear logs.

Enables MCP clients to observe Windows accessibility trees and deltas, generate fail-closed semantic action plans for external authorized executors, and control isolated, agent-owned Chrome tabs without accessing personal tabs.

Enables AI assistants to perform intent routing, retrieval reranking, and batch semantic judgments on Cloudflare Workers, returning structured answers with confidence scores instead of generated prose.

A unified remote terminal management server that simplifies complex SSH connections and server management through MCP protocol, allowing users to control multiple remote servers with natural language commands as easily as local operations.

A local MCP server that enables AI applications like Claude Desktop to securely access and work with Obsidian vaults, providing capabilities for reading notes, executing templates, and performing semantic searches.

MCP server for Windows desktop automation, enabling agents to control mouse, keyboard, and UI elements via 22 tools, with OCR, screenshots, macro recording, and autonomous mission execution.

Public MCP deliberation for AI agents: join, propose, argue, amend, vote, follow topics and invite peers through a Streamable HTTP endpoint. Humans can observe debates and conclusions; rules remain contestable and no model-provider API keys are requested.

Enables MCP clients to control the local Open Science research application, including creating projects, starting or continuing research sessions, inspecting progress and plans, canceling runs, and reading generated artifacts.

An open engine that gives AI persistent, reusable UI components with a writable component registry, versioned data storage, and a shell runtime, enabling AI to create, save, and reuse single-purpose apps.

Connects Claude Code and Codex to Jev-compatible System One services for context compaction, completion verification, and decision tasks like screening, classification, and gating.

Provides computational tools for systematic reasoning, including boolean evaluation, date arithmetic, object counting, state tracking, and format validation, to enhance agent capabilities and eliminate calculation errors in reasoning tasks.

Open Task Relay provides free, bounded public-good tasks that autonomous AI agents can discover, complete, submit, and independently verify. Its public TypeScript source runs locally with a Cloudflare Workers-compatible runtime and SQLite/D1, exposing MCP over HTTP.

Enables Claude Desktop to control mouse, keyboard, and clipboard on Windows, allowing AI-driven computer interaction tasks like clicking, typing, scrolling, and screenshotting.

An MCP server that lets Claude control a Minecraft bot with 40+ actions including movement, combat, crafting, and inventory management. Built on Mineflayer, it supports Microsoft authentication, pathfinding, and auto-reconnect.

Enables agents to control Windows desktop applications through UI Automation and Win32, supporting window management, mouse/keyboard input, element inspection, screenshots, OCR, macro recording, and autonomous multi-app missions. It also provides an HTTP/stdio MCP interface, an operator UI, and optional telemetry and safety controls.

Enables Claude Agent SDK agents to safely operate a remote user machine via a WebSocket tunnel, providing six client__* MCP tools with path fencing and phase-switch adjudication.

Model-agnostic computer-use tools over MCP, enabling screen capture and automated actions (click, type, scroll) via natural language tool calls.

Enables AI models to control a local Windows desktop by moving the cursor, clicking, scrolling, typing Unicode text, sending key chords, and capturing screenshots. Runs natively on Windows with minimal latency and no heavy dependencies such as Python, PyAutoGUI, or Docker.

A local MCP server that lets Hermes supervise Claude Code, delegating focused coding, research, or review tasks to the Claude Code CLI and managing worker sessions, background jobs, cancellations, and read-only reviews.

MCP server that gives any LLM a managed Docker workspace with live browser, terminal, code execution, document skills, and autonomous sub-agents.

Enables isolated shell command execution, file and process management, native image inspection, downstream tool-server proxying, persistent thread goals, and token-usage accounting through a live dashboard.

Lets AI agents see and control desktop applications through the accessibility layer, enabling clicking, typing, scrolling, dragging, and window/app management across macOS, Windows, and Linux entirely on the local machine.

Provides five rigorous reasoning protocols (debate, red team, audit_argument, threat_model, check_study) that run on the AI you're already using, requiring no extra API keys or costs.

Use Codex as a subagent inside Claude Code: async dispatch into a long-lived codex app-server, results pushed back into the Claude session automatically, steer/interrupt mid-task, and Codex's built-in code review.

Enables OpenAI Codex to inspect and operate Windows desktop applications in the background while the user keeps control of the mouse, keyboard, and foreground window, with each action verified against the application's real state.

Trust layer for AI agents: 22 pay-per-call x402 tools for scam and fraud checks, payment pre-flight, wallet screening, policy approvals, contact extraction and company research. Free preview with no wallet; paid calls settle in USDC on Base.

Transforms any prompt into a fully functional, production-ready product with zero human intervention by providing 150+ autonomous tools covering all aspects of software development.

A dynamic MCP server that automatically discovers Python files in a directory and exposes them as tools to any MCP-compatible AI client, allowing users to easily create and deploy custom AI tools.

Enables AI agents to control macOS via the ZCode CUA 0.5.12 MCP contract, providing accessibility-based app and window observation, screenshots, and input actions such as clicking, typing, scrolling, and clipboard access.