MCPViews

Give your AI agents a visual interface

MCPViews is an open-source, free-to-use companion that transforms plain-text tool output into interactive, beautifully rendered content — tables, diagrams, code reviews, and more. Through the Model Context Protocol, agents push rich content to a companion window.

Why MCPViews?

🔌

Plugin System

Install renderers for any data type. From database schemas to API docs, each plugin knows how to display its content.

🔄

Live API Hydration

Renderers call your backend API directly, fetching live data and enabling real-time interactions — blending MCP orchestration with native app UIs.

Rules Repository

Expose breadcrumbs, prompts, skills, and workflows through MCP so agents fetch rules only when a task needs them.

📊

Graphs & Data

Native graph packs, sortable tables, and raw data payloads rendered consistently without asking the agent to generate chart code.

Token Efficient

Agents can send compact data references while MCPViews handles formatting locally, reducing output tokens and making rendering more reliable.

Review Workflows

Accept or reject changes inline. Schema migrations, document edits, and code reviews all support interactive approval.

🎨

Theme Support

Dark and light modes that match your editor. MCPViews adapts to your workflow.

🔀

Bi-directional

User interactions in the companion window flow back to the agent as context for its next response.

How it works

AI Agent

Claude, Cursor, etc.

MCPViews

MCP Server

routes to

Plugin Renderers

Tables, diagrams, reviews

Companion Window

Visual output

hydrate

Your Backend API

Live data

Install MCPViews

Your device looks like an unknown operating system, so this guide shows only that setup path.

Step 1

Install MCPViews

Your operating system was not detected, so the macOS installer is shown by default.

Download MCPViews for macOS v0.2.6
  1. Open the downloaded installer.
  2. Move MCPViews into Applications if prompted.
  3. Launch MCPViews and leave the companion window open.

Step 2

Copy the setup prompt

After MCPViews is running, copy the prompt and paste it into your agent of choice, such as Claude Code, Codex, Cursor, or another MCP-capable agent. MCPViews requires that agent to support the Model Context Protocol.

Latest release metadata is loaded from GitHub releases.

Documentation

Loading...