"Information about modal concepts or elements" matching MCP servers:

Enables AI coding assistants to drive a real, persistent browser session through tools for goal-driven browsing, numbered screenshots and element lists, clicking, typing, selecting, scrolling, history navigation and setup checks. Each action is chosen by a cheap decision model, with a vision LLM stepping in only when that model is unsure.

Enables agents to get fast, calibrated probabilistic answers from Jev (Typesafe AI) to yes/no, scale, or choice questions about provided material, without using a generative model.

An improved implementation of persistent memory using a local knowledge graph with a customizable --memory-path. This lets Claude remember information about the user across chats.

A bridge that connects the Model Context Protocol to the Gemini CLI via the Agent Client Protocol (ACP) for high-performance, stateful AI interactions. It enables MCP clients to leverage Gemini's autonomous capabilities, including multi-modal support and built-in tool execution, without cold-start overhead.

A local stdio MCP server that exposes compact semantic observations from UI elements annotated with data-agent-* hints, designed to work alongside the official Playwright MCP for browser automation.

An MCP server that enables AI agents to pause and request human approval or information via Slack, Telegram, or macOS dialogs before proceeding with actions.

Enables complex problem-solving through multi-modal reasoning, dynamic knowledge graphs, meta-cognitive reflection, and collaborative AI personas.

Playwright for the entire OS. Give AI agents eyes and hands on any desktop app — find, click, type, and read UI elements across Linux, macOS, and Windows.

This MCP server enables AI agents to perform exploratory UI testing of live web apps by providing a structured view of page elements, deterministic correctness oracles, enforced read-only or opt-in write policies, persistent cross-run memory, and gap-checked coverage reports.

Windows desktop automation MCP server enabling screenshot capture, precise visual locating of UI elements via vision models, and automated mouse clicking.

Judgment calls on small, fast, cheap models: yes/no, pick-one and rating questions about any text, answered YES, NO or UNSURE. Four tools (likely, classify, rate, each) for TypeSafe's Jev, open models on Ollaya, Ollama or a local NLI model.

An MCP server that answers typed noul/choice/score questions about a given state with probability distributions and confidence, for routing, escalation, and gating decisions. It runs on pluggable backends (local Needle 3 or Laya models, or any OpenAI-compatible endpoint) and logs every decision so recorded ground truth can be used to track calibration.

Enables AI agents to analyze videos and images using Qwen3-VL deployed on Modal, supporting hours-long videos with timestamp grounding, text extraction, video summarization, and Q\&A with 256K context window.

A Model Context Protocol server that converts Markdown content into interactive mindmaps, allowing AI assistants to visualize hierarchical information through either HTML content or saved files.

Allows AI assistants to interact with Autodesk Revit through the MCP protocol, enabling the AI to create, modify, and delete elements in Revit projects.

Enables agent-native desktop control by driving mouse, keyboard, and interface elements through accessibility-first semantic operations, with fallback to screen coordinates when accessibility cannot reach a target.

Enables cleaning and sanitizing prompts through an LLM-powered tool that removes sensitive information, provides structured feedback with notes and risks, and normalizes prompt formatting. Supports configurable local or remote OpenAI-compatible APIs with automatic secret redaction.

Enables AI assistants to capture and analyze screenshots using Claude Vision API, providing AI-powered insights about desktop interface content, UI elements, and visual layouts.

Enables browser automation in which a planning LLM states one outcome at a time while a Typesafe System One model decides which element, action, and value to use, with Playwright carrying it out. Exposes tools for opening pages, driving multi-step goals, checking or choosing among options, reading relevant page text, acting on elements directly, and calling a page's own WebMCP tools.

Serves coding agents the annotations users leave on elements of a running dev app, including the element selector, DOM context, component name and exact source file:line. Agents can list and watch for pending feedback, link a page to their repository, and acknowledge, reply to, resolve or dismiss each annotation so the thread resolves on the page.