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.
-
Highly configurable YATT MCP server library: spin up your own UI-testing MCP server (35 tools, Playwright engine, headless-first) with strict, typed configuration for paths, HTTP transport, auth, permissions, sessions and retention.
npm MIT -
superbasedgithub.com/marmutapp/superbasedupdated this year
Screenshot capture, AI analysis, OCR, visual regression testing, token compression — headless server + MCP tools for AI coding CLIs
npm ISC -
Let your AI coding agent test its own frontend work. Connects Claude Code, Cursor, and any MCP client to your real Chrome browser — real sessions, real cookies, no headless conflicts.
npm MIT -
getjoinableactive
Joinable's self-test and MCP server: check a web game's multiplayer integration with two headless players, and give coding agents the Joinable kit. https://getjoinable.com
npm UNLICENSED -
MCP bridge for testing local open.mp / SA-MP servers through a headless RakClient.
MIT ★ 2 -
MCP server and CLI that makes Claude Code, Codex and any coding agent better at Godot 4: offline Godot 4.7 API, project map, headless checks, an AAA GDScript review, safe edits and renames, GUT/GdUnit4 tests, play-testing with input, screenshots, and 3D m
npm MIT -
mcp-unrealgithub.com/remiphilippe/mcp-unrealupdated this year
MCP server that gives AI coding agents (Claude Code, Cursor, etc.) full control over Unreal Engine 5.7 projects — headless builds & tests, Blueprint editing, actor manipulation, procedural mesh generation, and UE API documentation lookup.
Apache-2.0 ★ 82 -
Open-source Three.js game engine for TypeScript with Rapier physics, deterministic simulation, headless testing, and MCP tooling for AI agents.
npm MIT -
Stealth browser automation for AI agents — unattended, headless, N parallel persistent logged-in Chromes. One MCP server + CLI, credential vault with TOTP/IMAP 2FA, vision clicking, prompt-injection scanning. Self-hosted, real Chrome. 2,027 tests.
Apache-2.0 ★ 9 -
PicassoWeb MCP Server: An AI-agent tool for automated UI replication, CSS token extraction, and design system scraping via headless Chromium. Features pixel-differential testing, asset downloading, layout mapping, and component cloning.
MIT ★ 4 -
Point an app at two logged-in users and it finds the broken access control (IDOR). Deterministic browser perception + cross-identity authz testing for AI agents — drives your real Chrome over MCP or runs headless. Open-core, local, no account.
Apache-2.0 ★ 3 -
MCP server, CLI and Claude Code plugin that make Claude Code, Codex, Cursor and any coding agent better at Godot 4: offline Godot 4.7 API, AAA GDScript review, headless checks and tests, Godot 3 to 4 migration, and 3D models from code.
npm MIT ★ 2 -
Localhost-only headless browser MCP server for agent-driven QA of local dev servers
MIT -
MCP server that lets an AI coding agent run and drive a real GPUI desktop app: screenshots, element tree, clicks, typing, scrolling, dialogs and file pickers, headless or in a sandboxed KDE Plasma session on Linux
Apache-2.0 ★ 2 -
vwaygithub.com/chpr8/vwayactive
A virtual Wayland compositor an agent can use to test your desktop apps via MCP.
Apache-2.0 ★ 1 -
MCP server for Linux desktop GUI automation on KDE Plasma 6 Wayland — 29 tools for virtual and live sessions
MIT ★ 80 -
can-seegithub.com/hurleysk/can-seeupdated this year
MCP server that gives AI agents eyes on your terminal — PNG screenshots, visual diffs, GIF recording, and full keyboard/text interaction
npm MIT -
SeleniumBase for Rust (unofficial port): browser automation and testing with the SeleniumBase API, a driverless Chrome DevTools engine, stealth mode, and an MCP server for AI agents.
NOASSERTION ★ 1 -
periscope-mcpgithub.com/segentic-lab/periscope-mcpupdated this year
Web-app QA, testing & analysis for AI agents — 74 Playwright tools: authenticated flows (2FA/SSO), 25-action E2E workflows, hard assertions, session reports, and accessibility / SEO / GEO / Core-Web-Vitals + Lighthouse audits across a page or a full crawled site. Battle-tested by an AI agent on real apps.
AGPL-3.0 ★ 10 -
Autonomous headless browser QA & self-healing verification loop for coding agents. 8-trap deep inspection (console, network, layout, WCAG a11y + keyboard focus, Core Web Vitals, links, SEO, theme parity) across desktop/tablet/mobile in light + dark, QA He
npm MIT -
terminal-driver-mcpgithub.com/funkyfunc/terminal-driver-mcpupdated this year
A real interactive terminal for AI agents (MCP). Run and test vim, htop, gdb, REPLs, and any TUI — things the built-in fire-and-forget terminal can't touch.
npm MIT -
Three.js MCP server and Claude Code plugin: build three.js games and 3D pages end to end, from references and a design doc to playtesting, models and a published link. It tells the agent what is on screen, what never appears, what each input changed and w
npm Apache-2.0 -
🌟Zero-dependency MCP kernel for coding agents. Runs and tests speculative edits in RAM shadow workspaces before touching disk.
MIT ★ 1 -
Eyes, ears and hands for AI agents on GUI programs: a sealed X11 or Wayland session where an agent (over MCP), a script or CI reads the accessibility tree, takes screenshots, records video and sound, clicks, types and drives gamepads, then checks contrast, Tab order, visual diffs, flicker, input latency and audio dropouts. CLI, Python, Playwright.
npm MIT ★ 1 -
Autonomous QA scanner: point it at your running app and it walks the real user flows in a headless browser and reports what's actually broken: broken pages/links, console and network errors, empty/placeholder data, broken forms, state-loss, bad loading/er
npm Apache-2.0