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.
-
vlm-mcpgithub.com/example/vlm-mcpupdated this year
VLM Image Understanding MCP Server - Support OpenAI protocol compatible VLMs
pypi MIT -
MCP server for image analysis using the InternVL3 vision-language model
-
vlm-mcp-servergithub.com/syntx-ai/vlm-mcp-serverupdated this year
MCP server for vision and video analysis with pluggable VLM providers.
★ 2 -
The visual verifier for agentic coding workflows — deterministic-first UI screenshot diffing (DOM+pixel), tiered descriptions, CLI, MCP server, GitHub Action
npm MIT -
Document image retrieval via MCP or API for agentic systems using semantic embeddings, YOLO, and VLM classification.
NOASSERTION ★ 4 -
ros2-image-mcpupdated this year
MCP server (FastMCP) exposing ROS 2 image topics to VLM agents
pypi -
desktop-test-mcpupdated this year
MCP server for Windows desktop UI E2E testing with VLM-powered visual assertion
pypi MIT -
claude-kvm-nativegithub.com/remrearas/claude-kvm-nativeupdated this year
MCP server (Native) — control remote desktops via VNC, SSH and on-device VLM (macOS)
npm MIT -
LLM agent that finds objects in a simulated house — LangGraph + MCP + ROS 2 Nav2 + YOLO + VLM second opinion
MIT -
MCP-based laboratory vision robot: ESP32 voice terminal integration, cloud VLM image Q&A, ONVIF cameras and Opentrons.
MIT -
eyes-mcpgithub.com/JamesbbBriz/eyes-mcpupdated this year
👀 Give any text-only LLM eyes — local vision (VLM + OCR) for coding agents, via MCP. Zero API keys, nothing leaves your machine.
MIT -
MCP server exposing GroundingKit's VLM-based screen-region grounding to MCP-compatible AI agents. Native Swift, Apple Silicon, single ground_region tool.
MIT -
Give text-only LLM agents eyes: an MCP server + CLI that describes images through a vision-language model (VLM) API.
npm MIT -
MCP server for Wayland desktop automation. Screenshot capture with VLM analysis, pointer and keyboard control, and action chaining for any MCP client.
GPL-3.0 ★ 15 -
Captures d'écran → base de connaissances Markdown et Notion : VLM local, validation humaine, connecteur MCP. Vue 3 · FastAPI · Pydantic-AI · Docker. Projet personnel.
MIT -
Converters where figures survive. DOCX/XLSX to Markdown via native OOXML chart data: real numbers, OCR/VLM are only optional. CLI, MCP server, Docker image, and .mcpb bundle included.
pypi Apache-2.0 -
Visible AI-driven robotics demos powered by VLM policies, MCP skills, and AI coding agents. Reviewable HTML reports — frames, maps, tool traces, scores — instead of buried terminal logs.
MIT ★ 7 -
Hybrid local-first PDF parser in Rust (unrelated to ASP.NET RazorPDF): deterministic geometry for 90%+ of pages, selective sub-page VLM for the rest. Markdown + grounded RAG chunks via CLI, Python, MCP.
cargo MIT ★ 2 -
MCP server — 把 museav-cli(MUSE AV 出图中台)的能力暴露给 AI Agent:出图、抠图、超分放大、去水印、压缩、模板/技能/工作流/视频模板清单、模型与余额查询、素材上传、图生模板、SCULPT 反推、SkillHub 发布,外加 mlx-vlm-kit 本地看图理解(21 个工具,museav-cli ≥ 3.4.0)
npm MIT -
local_visiongithub.com/DavidEasden/local_visionupdated this year
An MCP that uses local VLM to convert images into descriptive text for large models without visual capabilities
Apache-2.0 -
Claude Code 的图像全家桶 MCP——AI 生图/视频 + 图像识别(OCR/表格/图表/VLM)+ 架构图/数据图/分享卡/二维码/公式/嵌套可下钻架构图,21 工具一个装好,纯免费。 | All-in-one image MCP for Claude Code — AI image/video gen + recognition (OCR/table/chart/VLM) + diagrams/charts/cards/QR/formula + nested drill-down architecture. 21 tools, free.
npm MIT ★ 2 -
pixel-mcpgithub.com/Todmy/pixel-mcpupdated this year
Figma-to-browser visual convergence harness — MCP server + CLI with escalating CV/DINOv2/VLM/human gates
Apache-2.0 ★ 1 -
search-and-vision-mcpgithub.com/dimthink/search-and-vision-mcpupdated this year
Give any LLM eyes and a search engine. Universal MCP server: web search (Tavily/Exa) + image understanding (any OpenAI-compatible VLM).
MIT -
vMLX - Use MLX models easily - JANGQ (GGUF for MLX) - Not dependant on mlx_vlm
pypi Apache-2.0 ★ 918 -
apudgithub.com/cadeos/apudactive
IIIF-grounded, citable vision for historical documents: region crops for VLMs, W3C annotations out. Python library + MCP server.
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 -
NoPPT : PPT = NoSQL : SQL. AI-native, Web-first HTML slide generation and editing platform with MCP support. Built-in LLM/VLM review engine, closed-loop feedback for auto-fix and regeneration.
MIT ★ 2 -
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 -
Desktop + browser GUI agent MCP server. VLM-powered automation: screenshot → vision model → mouse/keyboard/Playwright actions. 9 MCP tools. Fits RTX 4090.
Apache-2.0 ★ 2 -
LAN web UI + job runner for okforge knowledge bases: scanned PDF -> VLM OCR -> translated, page-cited LLM wiki. Serial job queue, MCP server, Quartz publishing.
MIT ★ 5 -
zelpiactive
ZelPi — Physical Intelligence Operating System. A self-contained CLI OS for autonomous robot fleets: kernel QoS router, model registry (VLM/VLA/LLM), MCP server, and an embedded simulation backend. Command: `npx zelpi`.
npm MIT -
Self-hosted VLM image-recognition MCP server (stdio). Paste an image in Claude Code / Codex / OpenCode → get recognition. OpenAI-compatible backend (vLLM / Ollama / GLM-4V / Qwen-VL).
MIT ★ 1 -
browser-use, but for the entire desktop — give any AI Agent eyes and hands on Windows. MCP + SDK, VLM-optional, with visual loop, introspection and meta-learning.
MIT ★ 14 -
verifydocgithub.com/bhaskargurram-ai/verifydocupdated this year
Trust layer for document→JSON extraction & AI agents: calibrated per-field confidence + source grounding + accept/review abstention on any OCR/VLM. Ships VerifyDocBench, a novel grounding-conditioned conformal method, and an MCP server.
Apache-2.0 ★ 1 -
gui-agent-frameworkgithub.com/sjkncs/gui-agent-frameworkupdated this year
Modular block-based GUI automation framework combining VLM vision, LLM planning, Playwright browser control, and pyautogui desktop automation. Includes food review crawler with multi-agent collaboration and MCP server.
-
pip install vidqa-cli — local, deterministic video QA CLI + MCP for tests, CI, and agents: 28 commands — frame timing, golden-frame diff, failure auto-locate, on-screen text index, color-aware run-to-run diff, Playwright/adb capture, local-VLM rubric review. One compact JSON + exit code per call. $0, fully on-device.
MIT ★ 1 -
@iflow-mcp/dcostenco-prism-mcp-servergithub.com/dcostenco/prism-mcpupdated this year
The Mind Palace for AI Agents — persistent memory (SQLite/Supabase), behavioral learning & IDE rules sync, multimodal VLM image captioning, pluggable LLM providers (OpenAI/Anthropic/Gemini/Ollama), OpenTelemetry distributed tracing, GDPR export, multi-age
npm MIT -
企业私有化知识库系统(单租户):文档摄取 → 混合检索 → 带引用的问答 → 内容治理与渠道接入。Docker Compose 一键部署,lite 单容器不需要 GPU、standard 五服务共用同一镜像。引用溯源到文档/章节/页码,命中图文章节附原图;扫描件 OCR 与 VLM 解析、表格结构保留(行组不拆散、跨页断表合并)、飞书 wiki 整树导入(含画板)与定时增量同步;Embedding/向量库/LLM/OCR/分块器全插槽;RBAC + 库级 ACL + SSO + 审计 + 评测回归 + 运营看板;OpenAI 兼容 API 与 MCP Server;飞书 / 企业微信机器人;中/英/马三语界面。通用定位,不绑单一客户场景。
★ 1 -
AI Scraping Stack — web intelligence platform for competitor analysis & automated scraping. Multi-agent pipeline extracts insights from URLs, analyzes screenshots with VLMs, generates production-ready HTML/React code. 10+ providers with cascade fallback, zero-cost mode with Ollama + OpenRouter. Built with FastAPI, Pydantic, MCP
MIT -
Preservation-aware visual dataset curation & contrastive safety alignment engine for AI. Replaces destructive whole-image sample-dropping and diffusion inpainting drift with deterministic sub-pixel Rust SDF masking, 540p JPEG survivability verification, and zero-glue MCP server for foundation VLM pre-training and DPO.
NOASSERTION ★ 1 -
AdaTile-MCP: high-resolution image adaptive tiling MCP server for DeepSeek vision model (deepseek-v4-flash-vision-exp). L1-L6 pipeline (fastpath, saliency, adaptive tiling, Files API assembly, streaming VLM, rule-based merge) + eval harness. Setup: clone, run setup.bat, add your agent MCP config -> see README.
MIT -
comfyui_LLM_partyygithub.com/heshengta0/comfyui_LLM_partyyupdated this year
LLM Agent Framework in ComfyUI includes MCP sever, Omost,GPT-sovits, ChatTTS,GOT-OCR2.0, and FLUX prompt nodes,access to Feishu,discord,and adapts to all llms with similar openai / aisuite interfaces, such as o1,ollama, gemini, grok, qwen, GLM, deepseek, kimi,doubao. Adapted to local llms, vlm, gguf such as llama-3.3 Janus-Pro, Linkage graphRAG
AGPL-3.0 ★ 2 -
comfyui_LLM_partygithub.com/heshengta0/comfyui_LLM_partyupdated this year
LLM Agent Framework in ComfyUI includes MCP sever, Omost,GPT-sovits, ChatTTS,GOT-OCR2.0, and FLUX prompt nodes,access to Feishu,discord,and adapts to all llms with similar openai / aisuite interfaces, such as o1,ollama, gemini, grok, qwen, GLM, deepseek, kimi,doubao. Adapted to local llms, vlm, gguf such as llama-3.3 Janus-Pro, Linkage graphRAG
AGPL-3.0 ★ 2 -
comfyui_LLM_partygithub.com/heshengtao/comfyui_LLM_partyupdated this year
LLM Agent Framework in ComfyUI includes MCP sever, Omost,GPT-sovits, ChatTTS,GOT-OCR2.0, and FLUX prompt nodes,access to Feishu,discord,and adapts to all llms with similar openai / aisuite interfaces, such as o1,ollama, gemini, grok, qwen, GLM, deepseek, kimi,doubao. Adapted to local llms, vlm, gguf such as llama-3.3 Janus-Pro, Linkage graphRAG
AGPL-3.0 ★ 2,390 -
DrawForgegithub.com/qweadzchn/DrawForgeupdated this year
—
MIT ★ 6 -
agent-vision-mcpgithub.com/idealizing/agent-vision-mcpupdated this year
Universal vision tools for AI agents via Model Context Protocol
pypi MIT -
mcp_imagesgithub.com/LiveBigOrange/mcp_imagesupdated this year
MCP 图片分析服务器 — 在 AI 工具中直接分析本地图片、剪贴板截图或 Base64 图片数据
MIT ★ 1 -
llm-advisor-mcpgithub.com/Daichi-Kudo/llm-advisor-mcpupdated this year
—
npm MIT ★ 2 -
page-vision-mcpgithub.com/666666999999666/page-vision-mcpupdated this year
Structured visual analysis pipeline for coding agents - give your AI agent the ability to see
pypi MIT -
lucida-vision-mcpgithub.com/dongyushenqi/lucida-vision-mcpupdated this year
通用视觉感知 MCP 服务器(眼睛不是大脑)/ General-purpose visual perception MCP server (eyes, not brain)
npm MIT ★ 1 -
mcp-recognize-imageupdated this year
MCP server that lets text-only models 'see' images via configurable vision LLM providers with priority fallback
pypi MIT -
MCP server that lets AI agents analyze, describe and read text in images through the free-first three.ws vision API.
NOASSERTION ★ 10 -
vision-audit-agentgithub.com/rogermsc/vision-audit-agentupdated this year
Turns photo and video from physical operations into structured, evidence-backed audit reports. Vision + agent engineering with a golden dataset, a CI regression gate, a coded failure taxonomy, and MCP exposure.
Apache-2.0 -
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 -
Multi-backend GGUF vision-language inference engine (moondream3.1 · VideoChat3 · Qwen2.5-VL · MOSS-VL · VideoMAE/ViViT · SmolVLM) with a unified CLI, Web UI and an MCP server for agents.
NOASSERTION ★ 4 -
antiquarIOgithub.com/JCarrascoF/antiquarIOupdated this year
A privacy first (local) document parsing tool and MCP server — convert PDFs and images to structured Markdown using multiple selectable OCR engines. Includes Streamlit GUI and hierarchical chunking for RAG workflows.
MIT