MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
codictor-mcpupdated this year
Screenshot to code MCP that uses your project's design system
npm MIT -
blip-mcpgithub.com/nebenzu/blipupdated this year
Visual annotation MCP server for Claude Code - draw on preview screenshots to communicate UI changes
npm MIT -
glimpse-mcpupdated this year
MCP server that gives AI coding agents visual access to frontend via screenshots, diffs, DOM inspection and console logs
npm MIT -
screen-capture-mcpgithub.com/kmoulder/screen-capture-mcpupdated this year
MCP server that gives Claude Code the ability to take screenshots (Windows)
npm MIT ★ 8 -
MCP server for Hostbeam: lets Claude Code, Codex or any MCP client read the screenshots you beam from your Mac to the server it runs on, with nothing to paste. npx -y @hostbeam/mcp
npm MIT -
@framesnap/mcp-serverupdated this year
MCP server for FrameSnap — give AI coding agents the ability to take screenshots
npm MIT -
pageshot-mcpgithub.com/softvoyagers/pageshot-apiupdated this year
MCP server for PageShot — take webpage screenshots, render HTML to images, and generate PDFs from AI assistants like Claude, Cursor, VS Code
npm MIT -
Annotate your screen and feed it to your AI agent. Press a global hotkey, circle bugs, draw arrows, type labels; the annotated screenshot and structured metadata drop into an inbox that Claude Code (or any MCP agent) can read and act on.
MIT ★ 11 -
contextshotupdated this year
MCP server for ContextShot. Exposes iPad drawings, screenshots, and screen recordings to AI tools like Claude Code and Cursor.
npm MIT -
MCP server for the Zactonz APIs: screenshots, page to Markdown, link previews, DNS, SSL, WHOIS, email checks, QR codes and more as tools for AI assistants.
npm MIT -
MCP server that teaches AI coding assistants how to author visual/screenshot regression tests for the storybook-addon-playwright Storybook addon
npm MIT -
Pin feedback to the live UI. Hand it to Claude. An embeddable visual-feedback SDK — inspect an element, pin a comment, capture a screenshot; comments re-anchor across redeploys and reach Claude Code via MCP.
npm MIT ★ 2 -
BD-VisionShotgithub.com/FDiver/BD-VisionShotupdated this year
Lightweight Windows screenshot tool for Claude Desktop / Claude Code + agent-vision-mcp (MCP). Capture a region -> copy local PNG path + prompt to clipboard -> paste into Claude -> MCP reads and analyzes the local image. Global hotkey, tray, single-file EXE, fully local, no uploads. 截图 识图 剪贴板 工具.
MIT ★ 1 -
MCP server that hands your Notepix browser annotations (and their screenshots) straight to an AI coding agent.
npm MIT -
alioss-dropgithub.com/karrychow/alioss-dropupdated this year
Upload images/screenshots to Aliyun OSS and get public URLs — single Go binary, integrates with Claude Code via MCP
MIT -
Give your AI coding agent access to Figma design data. 9 tools covering design context, variables, screenshots, metadata, FigJam, and more.
npm MIT -
autopilot-mcpactive
MCP server wrapper for Autopilot — exposes ax_navigate, ax_screenshot, ax_accessibility, ax_execute, ax_wait, ax_text to Claude Code
npm MIT -
MCP server for comparing UI screenshots using pixelmatch to help autonomous coding agents validate frontend implementations.
MIT ★ 1 -
@glenmaura/autopilot-mcpactive
MCP server wrapper for Autopilot — exposes ax_navigate, ax_screenshot, ax_accessibility, ax_execute, ax_wait, ax_text to Claude Code
npm MIT -
@hyzer-labs/annotategithub.com/hyzerlabs/annotateupdated this year
Annotate any page in the browser and send the element, screenshot, and your comment to your local coding agent over MCP.
npm GPL-3.0-only -
Drive Claude Code from Telegram: code from your phone, approve every command with a tap, send screenshots. Plus an MCP server to message yourself back.
MIT -
MCP server exposing browser flows recorded by the FlowSnap Chrome extension — steps, console errors, network calls and screenshots — to Claude Code.
npm MIT -
MCP server exposing browser flows recorded by the DevFlow Chrome extension — steps, console errors, network calls and screenshots — to Claude Code.
npm MIT -
MCP server that lets Claude Code drive App Store Connect end-to-end (build, upload, metadata, screenshots, submit for review).
MIT ★ 10 -
@pinagent/next-plugingithub.com/pinagent/pinagentupdated this year
Next.js plugin for Pinagent — capture in-browser UI feedback with file:line and a screenshot, hand it to your coding agent over MCP.
npm Apache-2.0 -
Point-and-click website annotation that hands small UI fixes straight to a coding agent. Annotate in a real browser, let the agent fix, verify with before/after screenshots.
AGPL-3.0 ★ 1 -
MCP server exposing read-only Mozaic ADS2 Vue design-system documentation (components, icons, styles) and the screenshot-to-code workflow to AI agents.
npm MIT -
Local, click-to-comment visual review for HTML you're still building — Claude Code publishes a block to a local review server, you click an element and leave feedback, Claude reads it back anchored by a screenshot.
npm MIT -
Allows AI agents (e.g., Cursor, Claude Code) to debug within Tauri apps via screenshot capture, window management, DOM access, and simulated user inputs.
npm ★ 113 -
Model Context Protocol server for visual feedback and UI testing automation. Take screenshots, compare against reference designs, and get actionable feedback to help coding agents iterate faster on UI development.
MIT -
cursor-browser-bridgegithub.com/VectorlyApp/cursor-browser-bridgeupdated this year
Bridge Cursor IDE's built-in browser to any MCP client (Claude Code, Codex, etc.). Works over SSH. Navigate, click, type, screenshot, and more.
★ 2 -
Connects AI assistants like Codex, Claude and Cursor to Godot 4 to create and edit scenes, run games, debug code and capture screenshots.
npm MIT ★ 5 -
Xenocept plugin for Anthropic Claude — delivers sessions to Claude Code via MCP channels or polling, AND optionally runs Claude as the OCR / vision provider for screenshot enrichment. One Anthropic API key, two roles.
npm MIT -
A Claude MCP server to reverse engineer any feature like a pro hacker. Paste any link or screenshot and get production-ready code instantly.
-
Ignaru Studio from your terminal: give your coding agent a design canvas and taste. `npx -y ignarucli mcp --root .` serves the canvas, the design rules and skills, real fonts and screenshots over MCP; `ignaru connect <agent>` adds it to Claude Code, Codex
npm MIT -
MCP server that exposes a PC's capabilities (shell, filesystem, GPU, screenshots, background processes) to Claude Code or any MCP client
pypi MIT ★ 9 -
@mcpity/mcpgithub.com/mcpity/mcpityupdated this year
MCP server for MCPity Studio — push code to engines, screenshot viewport, render video, deploy games, build and ship SaaS websites
npm MIT -
@smile-design/mcp-servergithub.com/smilezyl2023/smile-designupdated this year
MCP Server for smile-design — exposes design tokens, aesthetic checks, DOM inspector, preview lifecycle, and screenshot tools to Claude Code so any IDE / CLI agent supporting Model Context Protocol can drive smile-design's design intelligence.
npm MIT -
viewgraphgithub.com/sourjya/viewgraphupdated this year
ViewGraph is a developer super assistant that captures live web pages as structured, agent-ready UI context, combining layout, DOM-like structure, selectors, styles, spatial relationships, screenshots, and metadata to help coding agents generate tests, audit interfaces, compare changes, and build faster.
npm AGPL-3.0 ★ 7 -
gecko-mcpgithub.com/Frumane/gecko-mcpupdated this year
MCP server to drive Firefox-based browsers (Floorp, LibreWolf, Waterfox, Zen, Mullvad, Firefox) from Claude Code, Cursor & any MCP client - read pages, screenshot, automate tabs & forms in your real session. 'Claude in Chrome', but for Gecko.
npm MIT ★ 2 -
Give your AI coding agent the real numbers behind any web page: sizes, spacing, type, colors and the animation timings a screenshot can't carry. Bridges the Design Toolkit browser extension to Claude Code, Cursor, Windsurf and any MCP client.
npm MIT -
MCP server for Android — drive emulators and real devices over adb from Claude Code, Cursor, or VS Code. 73 tools: screenshots, UI hierarchy, tap/swipe/type, logcat, device locks, Gradle builds and tests. The Android counterpart to XcodeBuildMCP.
go MIT ★ 20 -
Click elements on your live Vite dev page, write what should change, and send exact JSX source locations (file:line:column) and element screenshots straight to Claude Code or OpenAI Codex via MCP, plugin, or clipboard.
npm MIT -
CodeRadargithub.com/officialCodeWork/CodeRadarupdated this year
React & TypeScript static analysis that maps every UI component to the APIs, state, and events behind it — and serves that context to AI coding agents over MCP. Screenshot or ticket in; component, data lineage, user journeys, and blast radius out. Deterministic, no LLM in the core.
MIT -
MCP server bridging AI coding agents to a real Chrome browser via the AgentBridgeLens extension — DOM inspection, screenshots, network/console capture, source-map tracing, visual overlays, multi-tab targeting.
npm MIT -
agent-vision-mcpgithub.com/kitlau86/agent-vision-mcpupdated this year
An MCP server that gives non-vision LLMs the ability to "see" images. Plug in any OpenAI-compatible vision API (Gemini, Qwen-VL, OpenAI, or self-hosted) and your text-only model — like DeepSeek in Claude Code — can analyze screenshots, OCR text, read charts, and more.
MIT ★ 10 -
luma-mcpgithub.com/JochenYang/luma-mcpupdated this year
多模型视觉理解 MCP 服务器,为不支持图片理解的 AI 编码模型提供视觉能力:分析截图、报错、UI 与文档,可接入多家主流视觉大模型。Multi-model vision MCP server that adds image understanding to AI coding models without native vision — analyze screenshots, errors, UI and documents via major vision LLM providers.
npm MIT ★ 116 -
scrapepass-mcpactive
MCP server that connects AI apps on your computer (Claude Desktop, Claude Code, Cursor …) to the ScrapePass browser extension: extract lists and pages, magic fields, bulk downloads, screenshots, browser action flows, recipes, templates and saved data.
npm UNLICENSED -
Apixies gives Claude 50 read-only developer tools through the Apixies API. Inspect SSL certificates, DNS, WHOIS, security headers, redirects, sitemaps and page speed. Convert JSON, CSV, YAML, XML, Markdown and HTML to PDF. Generate UUIDs, hashes, passwords, QR codes and screenshots. Needs a free API key.
MIT -
Shows what your iOS app is actually doing — the screen, the app's own state, its network traffic and its logs — to an AI coding agent on your Mac, so checking a UI change stops being a manual tap-and-screenshot loop.
MIT -
Computer use MCP server for Claude Code, Codex and other AI agents. Reads the UI tree instead of taking a screenshot every step, with Jev (TypeSafe System One) as a fast judge that can also look at the screen to find icons and game objects. Local macOS, local Windows, or remote Windows over SSH.
MIT ★ 4 -
Showcase of an AI copilot for Db2 incident investigations: it reads insurance data through a read-only MCP server, links every finding to evidence, and leaves the decision to the investigator. Demo video, screenshots and architecture decisions. Synthetic data; source code private.
-
eyesongithub.com/grzetich/eyesonupdated this year
Comprehensive UX analysis platform powered by Google Gemini AI. Captures multi-viewport screenshots, identifies accessibility issues, detects AI design anti-patterns, and generates ready-to-implement code fixes. Available as web dashboard, CLI tool, and natural language interface via MCP.
-
one-shot-uigithub.com/TN0123/one-shot-uiupdated this year
Deterministic screenshot diffing for AI coding agents. Extract design tokens, diff implementations vs reference, get CSS fix suggestions.
npm MIT ★ 6 -
claude-code-gemini-ui-assistant-mcpgithub.com/lightwater2/claude-code-gemini-ui-assistant-mcpupdated this year
MCP server for Claude Code - collaborative UI generation: Claude orchestrates, Gemini generates components
npm MIT -
componentizegithub.com/MonloHua/componentizeupdated this year
截图→可复用的Tailwind组件。MCP Server,支持Claude/Cursor/Windsurf/Copilot等AI工具
MIT ★ 2 -
Professional Figma Component Restoration Kit - MCP tools with snapDOM-powered high-quality screenshots, intelligent shadow detection, and smart debugging for Vue component restoration. Includes figma_compare and snapdom_screenshot tools.
npm MIT -
devtools-mcp-servergithub.com/command-sector/devtools-mcp-serverupdated this year
MCP server providing 70+ developer tools - QR codes, PDFs, screenshots, crypto prices, email validation, AI text processing, and more. Use with Claude Code, Cursor, Windsurf, and other AI coding assistants.
npm MIT -
MCP server for LLMs to take screenshots on Wayland and X11.
-
tauri-mcp-servergithub.com/hyper-builders/mcp-server-tauriupdated this year
A Model Context Protocol server for Tauri v2 applications with screenshot-to-file support
npm MIT