MCP servers · shortlist
MCP servers for browser and OS automation
Browser drivers, desktop control, smart-home bridges and embedded devices — servers that let an agent act on a machine instead of only reading from it.
How this list was ordered. There is no reliable popularity signal in the source registries, so entries are ranked by how completely the record is documented — a substantial description plus a documentation, repository or base URL — and then alphabetically. 115 mcp servers in this scope clear the content bar; the top 24 are shown. A place here is not an endorsement or a paid slot.
Shortlisted entries
- 01
DevHusnainAi/zerodom
Deterministic attack-surface perception for AI security agents. Attaches over chrome.debugger to the Chrome you're already logged into (cookies, MFA and SSO already done) and turns the live page into a short numbered interaction graph an agent can act on ([03] button 'Sign In'), with CSS selectors kept server-side and out of the model's context. Cross-identity tools (compareidentities, replay) flag a byte-identical response across two logged-in users as a cross-tenant IDOR, and scan maps the attack surface from the CLI. No LLM in the parse; 71% fewer tokens than Playwright's ARIA snapshot mode.
Browser Automation auth: API Key
- 02
AbhijatSaxena/blink-camera-mcp
Control Amazon Blink cameras and their pan/tilt mount: read where the camera is pointing and its travel limits, aim it by absolute or relative angle with the new position confirmed by the hardware, return it home, and grab a snapshot. The mount's command channel is undocumented, so it was recovered from the Blink app's bytecode and verified against a real camera. uvx blink-camera-mcp
Home Automation auth: None
- 03
magnusmalm/smolmux
Serial and GDB/SWD MCP servers backed by a broker that shares one device link between agents, a human monitor and scripts: console history, send/expect, boot-stage tracking, U-Boot autoboot break-in, flasher handoff, breakpoints, Cortex-M fault decode, and unknown-board probing (chip identify to starter profile).
Embedded System auth: None
- 04
morfiade-app/morfiade-mcp
Manage local Chrome profiles, each with its own cookies, sessions and proxy: create, launch and stop them, assign and check proxies, tags, notes and trash with restore. Only whitelisted profiles are visible and there is no cookie-reading tool. Requires the Morfiade desktop app (Windows).
Browser Automation auth: None
- 05
Mohabdo21/linux-mcp
A Linux system monitoring MCP server providing real-time CPU, memory, disk, network, processes, Docker, systemd, security audit, and hardware info via STDIO transport. Install: download from GitHub Releases or go install github.com/Mohabdo21/linux-mcp@latest.
OS Automation auth: None
- 06
Savvy-Tech-Sphere/usable-browser-agent-free
MCP server plus Firefox add-on that lets your agent work in your real, logged-in Firefox: accessibility snapshots with stable element refs, logins filled by alias from a local vault, hardened mode by default. Free for personal use.
Browser Automation auth: None
- 07
SnapBlock/hai-browser
Let coding agents drive VS Code's Integrated Browser alongside you: snapshots, mouse and keyboard, forms, tabs, screenshots, console and network logs, a visible agent cursor, and click-to-source for React (Vite, Next.js).
Browser Automation auth: None
- 08
IGELHARINE/CDPM
Numbered, logged-in Chrome slots (1-30) driven over raw CDP that several Claude Code sessions and subagents share, with per-tab command queues, agent-to-agent messages, and recovery after reboots and Chrome updates.
Browser Automation auth: None
- 09
Metadrama/obscura-mcp
MCP server adapter for the lightweight Rust headless browser Obscura — high-performance web scraping with anti-detection. Perfect for AI agent automation. Server can run locally or as hosted endpoint.
Browser Automation auth: None
- 10
Eric-huang799/duplex
A browser shared by a human and an AI — same tabs, same live session; 24 MCP tools (snapshot, click, type, screenshot, tabs, annotations), built-in models or any MCP client, Esc takeover.
Browser Automation auth: None
- 11
emirmrkaljevic/usb-rea
Reverse engineer USB gadgets: capture vendor-app traffic (usbmon/USBPcap), infer the protocol with frame-level evidence and generate a pyusb driver. Replay requires human confirmation.
Embedded System auth: None
- 12
Miladmet/quickcast-mcp
Screen-to-Action Protocol: Turn screen recordings directly into automated Playwright E2E reproduction tests, structured GitHub bug reports with video timestamps, and SOP walkthroughs.
Browser Automation auth: None
- 13
tech-lead-bunny/agenteyes-mcp
Give coding agents real eyes: one tool call returns a screenshot PNG plus the accessibility tree and page text; segmentheight paginates long pages into readable tiles.
Browser Automation auth: None
- 14
faze79/WPFVisualTreeMcp
#⃣ - Inspect, debug and drive running WPF apps without source changes: visual tree, dependency properties, bindings, DataContext, screenshots, clicks and text input.
OS Automation auth: None
- 15
dockndevai/mcp-mac-control
Full Mac control: shell, AppleScript, files, processes and human-like GUI input with a screenshot loop, plus an optional safe mode with read-only, gated access.
OS Automation auth: None
- 16
Harusame64/desktop-touch-mcp
Windows desktop automation with entity-based actions instead of coordinate clicks, using UIA, CDP, screenshots, keyboard, mouse, clipboard and terminal control.
OS Automation auth: None
- 17
juergenkoller-software/nemeton-mcp
Control Nemeton, a macOS virtual machine manager built on Apple's Virtualization framework: Linux and macOS VMs, APFS snapshots, console, files and networking.
OS Automation auth: None
- 18
andresolbach/nodriver-mcp-server
Undetected Chrome automation via nodriver for anti-bot protected sites: accessibility snapshots, input automation, network inspection and persistent sessions.
Browser Automation auth: None
- 19
fixed-width/glass
Drive native GUI apps under development: launch, screenshot, read the accessibility tree, inject input, tail logs and diff frames on desktop, Android and iOS.
OS Automation auth: None
- 20
munimtechnologies/munim-computer-use
Accessibility-first computer use on macOS, Windows and Linux: element IDs instead of pixels, background input, a pointer overlay and its own Chrome tab group.
OS Automation auth: None
- 21
yunfeng-enrich/usemymac
Turn a Mac into a remote computer-use server via CUA: screenshots, mouse, keyboard, windows and apps, token-protected and exposed through a Cloudflare tunnel.
OS Automation auth: API Key
- 22
chryaner/terrarium
Fork, break and revert VirtualBox VMs with linked clones and RAM snapshots, and drive GUI-only guests via screenshots and input without SSH or a guest agent.
OS Automation auth: None
- 23
Cubenest/rrweb-stack
Record your authenticated browser session (rrweb DOM, console, network) to local SQLite with PII masking, so coding agents can query it or drive the browser.
Browser Automation auth: None
- 24
glasswarp/mcp-server
See and control your own Windows PC locally or remotely: UIA and screenshot observation, clicking, typing, dragging, scrolling, launching apps and Live View.
OS Automation auth: None
Snapshot 2026-10-19. Fields shown are what the source registry stated, not a live probe — check the linked documentation before integrating. See the methodology.