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.
-
perception-mcpgithub.com/egorthinks/perception-mcpupdated this year
Perception & verification MCP server for AI agents: media_assert (declarative pass/fail QC for generated media), image diff (pixelmatch + SSIM), readable text tiles for small print & handwriting, video tools (frames on demand, scenes, QC defects, pacing dynamics, attention heatmaps, scene diff), and full audio DSP (pitch, BPM, key, LUFS, spectrograms).
pypi MIT License Copyright (c) 2026 Egor Fedorov Pe -
hollow-mcpupdated this year
MCP server for Hollow — serverless web perception for AI agents
npm Apache-2.0 -
Measured visual perception for AI agents: OCR, geometry, colour and version diffing over MCP.
pypi Apache-2.0 -
cortexgithub.com/tsubasa-rsrch/cortexupdated this year
Cognitive-science-based perception framework for AI agents. Habituation, circadian rhythm, scheduling, notifications — stdlib-only, zero dependencies.
MIT ★ 3 -
mcp-browser-agentgithub.com/akshay2604/mcp-browser-agentupdated this year
MCP server giving AI agents structured perception of web pages — semantic UI tree, actions, and linting via Playwright + CDP
npm MIT -
afferens-mcpgithub.com/afferens/afferens-mcpupdated this year
MCP server for Afferens — real-time physical perception for AI agents
npm -
Perception layer for AI agents — fuses visual + structural + temporal data into a UI Scene Graph. Exposed as an MCP server.
★ 1 -
app-screen-mcpgithub.com/xmuweili/app-screen-mcpupdated this year
MCP server for iOS Simulator automation — accessibility tree, screenshot, tap/swipe/type, and AI-friendly screen perception tools
npm ★ 2 -
Eyes & hands for AI agents: a desktop-automation MCP server (screen perception + input control).
cargo NOASSERTION -
AI-powered document perception and analysis MCP server with intelligent provider selection
pypi Apache-2.0 -
audientgithub.com/es617/audientupdated this year
Audio perception layer for AI agents — MCP server with continuous-listening detection and a concept memory that grows through use.
pypi Apache-2.0 ★ 1 -
AI smart home assistant for Home Assistant: LangGraph agent + MCP tools + RAG knowledge graph + vision perception
MIT ★ 17 -
agent-lensgithub.com/JakeRoyRandall/agent-lensupdated this year
Chrome extension + MCP server that gives AI agents enhanced browser perception — DOM diffing, annotated screenshots, page readiness, viewport context, extension management, and semantic regions.
-
Token-efficient AI browser bridge. MCP server + Chrome extension for structured browser perception and control.
npm MIT ★ 1 -
glazyrvizgithub.com/mcpmessenger/glazyr-vizupdated this year
Zero-Copy MCP Vision Server — sub-8ms browser perception for AI agents. Connect at https://mcp.glazyr.com/mcp/sse
npm -
thunder-eyegithub.com/orellius/thunder-eyeupdated this year
AI gets eyes into desktop apps — MCP server for visual perception of Electron, Tauri, Flutter, Qt, and native applications
npm MIT -
Physical camera perception, measurement, and verification tool for MCP AI agents
pypi MIT -
A perception and attention layer for AI agents: append-only Observations in, permission-filtered world views and evidence-backed Signals out
MIT -
aseprite-mcpgithub.com/bachhoang0606/aseprite-mcpupdated this year
Live-first Aseprite MCP server — drive a running Aseprite session from an AI agent (perception, constrained colour, tilemap tooling).
MIT ★ 4 -
cleyagithub.com/severment/cleyaupdated this year
Creative-generation and video toolkit for AI agents — declarative scene DSL, deterministic HTML→MP4 rendering, and perception instruments that measure the output
Apache-2.0 -
A browser body for an AI agent: drive a real, persistent Chrome over MCP. Accessibility-tree perception, act-by-name, no CDP.
MIT ★ 1 -
Motion perception for AI agents: see and measure video, animation and synthetic scenes as labelled frames and real numbers. CLI, MCP server and TypeScript library.
MIT -
Real-time Windows computer use for AI agents over MCP: screen perception, bounded reflexes, closed-loop motor control, and experimental fly-connectome controllers.
MIT -
A token-saving MCP server for Windows. Feeds AI agents structured UI Automation state instead of screenshots — ~94-98% fewer perception tokens per action, measured.
npm MIT ★ 1 -
ios-agent-drivergithub.com/CodeJonesW/ios-agent-driverupdated this year
MCP server that lets an AI agent drive the iOS Simulator in a loop — accessibility-tree-first perception with screenshot fallback, built on simctl + idb.
MIT -
Audio perception for AI agents: measure, see, compare, edit and synthesize sound over MCP. Loudness, spectrograms, pitch, events, transcription, cleanup.
pypi MIT -
MCP server letting AI agents drive a Linux/X11 desktop — AT-SPI accessibility-tree perception with element-level actions, instead of screenshot + coordinate clicking.
MIT -
Browser perception for AI agents — framework state, DOM, pixel-accurate text, network, screenshots, OCR, and a UI state graph — exposed over MCP and HTTP.
npm MIT ★ 1 -
The browser your AI pilots — an open-source, AI-native browser with an autonomous copilot. Pure-CDP engine (no Playwright), token-first perception, CLI + MCP.
GPL-3.0 -
The AI's hands and eyes on the web - an AI-native browser framework with structured perception, verified self-healing actions, and an MCP server.
npm MIT ★ 1 -
Give any AI agent your real Chrome: token-cheap text perception, a visible cursor, MCP + HTTP API. Optional chat watcher.
npm MIT ★ 3 -
Peripheral Evaluation & Execution Proxy: The open-source token shield for AI coding harnesses that offloads mobile visual perception, coordinate grounding, micro-loops, and noisy log analysis to local models.
npm Apache-2.0 -
A high-speed, zero-cloud OS perception layer and MCP server for Windows. Delivers sub-15ms semantic focus, tab discovery, and workspace telemetry for local AI agents and voice interfaces
NOASSERTION ★ 1 -
OS-Native Windows Computer-Use Engine, CLI & MCP Server (54 Tools) in Pure Rust. Sub-millisecond Set-of-Marks UIA perception, atomic action pipelines, vector canvas engine & Per-Monitor DPI v2 input for AI Agents.
MIT -
windows-computer-use-mcpgithub.com/betrayzl/windows-computer-use-mcpupdated this year
Human-like Windows control for AI agents via the Model Context Protocol. Combines DXGI visual perception with simulated mouse/keyboard input — works on any Windows application without API integration. Built by a non-programmer with Claude Code.
npm MIT ★ 5 -
Give any AI agent general-purpose computer use: live screen perception (OCR, vision frames, MJPEG stream) plus safety-gated mouse, keyboard, window, virtual-desktop and game-mode control over a local HTTP API and MCP.
MIT ★ 10 -
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.
npm Apache-2.0 ★ 3 -
Percept — Ambient Voice Intelligenceno update data
Give your AI agent ears — ambient voice intelligence with 8 MCP tools for Claude Desktop
pypi -
Delegated vision MCP server for local images and webpage screenshots.
Apache-2.0 -
Give your AI agent eyes on the physical world. MCP server for real-time DePIN edge-vision queries.
MIT ★ 1 -
axiom-perception-mcpgithub.com/vdalhambra/axiom-perception-mcpupdated this year
Empirical memory for AI agents — patterns learned from real executions, not hand-written rules. Workflow checkpointing, failure knowledge base, multi-agent coordination. Zero API keys required.
pypi MIT -
Let any coding agent watch a video: local scenes, transcript and frames, with every answer pinned to a timestamp.
MIT -
perceptgithub.com/wowow4978-vorfeed/perceptupdated this year
A local service that aggregates heterogeneous sensor data into a uniform, LLM-queryable event store. An LLM can ask "what do you see now?" or "what happened last night?" and get grounded answers from real sensor history.
MIT -
Ollama for vision: pull, run & serve computer-vision models (YOLO, SAM 2, Depth Anything, Grounding DINO) from your terminal. REST, WebSocket & MCP for AI agents and robots.
MIT ★ 1 -
EDNA — Human Sensoric for AI. Behavioral Biometric Human Sensoric Engine: senses 16 channels of human behavior and makes them AI-readable, 100% on-device. MCP server, API, SDK, PWA. Independently verifiable.
npm SEE LICENSE IN LICENSE -
Agentic retail shelf inspection: real OpenCV 5 computer vision finds empty-shelf regions with quantified uncertainty, and a bounded Amazon Bedrock agent loop (via restricted MCP tools) actively decides what to inspect next. Deployed on AWS with a single CloudFormation template.
MIT -
manasgithub.com/ninthhousestudios/manasupdated this year
Modular agent infrastructure in Rust — memory, perception, coordination, planning, code intelligence for LLM workflows
AGPL-3.0 ★ 1 -
Simple RAG implementation from scratch using MCP, focusing on Perception, Memory, Decision and Action
Apache-2.0 ★ 15 -
mentu-mcpgithub.com/mentu-ai/mentu-mcpupdated this year
MCP intelligence layer — intent routing, perception, judgment, and sandboxed code execution across child MCP servers.
npm Apache-2.0 -
Perceptron MCP server for high-accuracy visual perception powered by fast, efficient vision-language models
npm Apache-2.0 -
World perception system combining computer vision models and systems like YOLO, SLAM, Segmentation, VLMs etc into a unified World State Summarizer Interface MCP
MIT ★ 5 -
Computer Experience Layer — MCP server for native computer use on macOS. Structured perception, reliable actions, continuous awareness.
npm Apache-2.0 ★ 6 -
Bridges the "black box" problem by creating a formal, inspectable decision layer. Each decision traces to explicit logical operations - no hidden states. Leverage NN perception while maintaining human-interpretable decision logic.
MIT ★ 2 -
Give Jev eyes — honest, local image perception for TypeSafe's text-only System One model. OCR + spatial layout → Jev state. CLI, MCP server, agent skill.
MIT ★ 1 -
cutiblegithub.com/plokdalberb-byte/cutibleupdated this year
Agent-native montage engine: Timeline-as-Data, verb API, deterministic FFmpeg render, QC loop, MCP server, ingest pipeline, multi-agent swarm, VLM perception, Remotion, OTIO bridge, render farm, REST API, SDK.
MIT ★ 2 -
craftonomousgithub.com/ReverseZoom2151/craftonomousupdated this year
Craftonomous is an agent-agnostic Minecraft embodiment and evaluation substrate: an MCP-native body with a declared perception budget and reliability-tracked skills.
MIT -
MCP server that drives a rooted Android phone from OS-reported controls: semantic perception, a resident native input daemon, and probe-and-remember learning. No OCR, no fabricated labels.
MIT ★ 4 -
organoid-mcpgithub.com/tan412/new_organoidupdated this year
MCP stdio shell for the organoid embodied-robot dataset QA kernel: 9-format lossless ingestion, 10 validators, six-state claim ledger. Perception-grade honesty: not_evaluated means insufficient evidence, never failure.
npm MIT -
An MCP server that gives LLMs full web-browser capabilities: navigation, interaction, JavaScript execution, vision/perception tools, network interception, and stealth features (fingerprint spoofing, humanized input) — including captcha detection and handling. Ships with Docker support and HTTP client examples.
NOASSERTION -
MCP server for Windows desktop automation. Incremental perception instead of a screenshot per step: measured 16.9x fewer tokens per action than the alternatives, with the benchmarks to check it.
pypi MIT ★ 4