Screeny MCP Server

Overview

The Screeny MCP server delivers a privacy‑first, macOS‑only solution for AI assistants that need visual context. By exposing only pre‑approved application windows, it guarantees that screenshots are taken without revealing unintended content or interrupting the user’s workflow. This design is especially valuable for developers who rely on AI agents to debug UI issues, generate documentation, or automate visual testing while maintaining strict privacy controls.

Screeny’s core functionality is centered around two lightweight tools: and . The former returns a curated list of windows that the user has explicitly granted permission to capture, ensuring no accidental exposure. The latter captures a high‑fidelity JPEG image of the specified window in the background, without requiring focus or visibility changes. This non‑intrusive capture is ideal for continuous integration pipelines and live debugging sessions where the AI agent must observe application state without disrupting user activity.

Key capabilities include configurable JPEG compression (default medium at 250 KB, adjustable between 100–900 KB), automatic deletion of image data after transmission, and a dedicated resource that provides server metadata. The server runs entirely locally, eliminating any external network connections and reinforcing its privacy posture. Developers can integrate Screeny into existing MCP workflows by adding a single server configuration, after which Claude Desktop or Cursor can invoke the tools directly from prompts.

Real‑world use cases span automated UI testing, where an AI agent can capture the state of a dialog after a test step; documentation generation, allowing agents to embed screenshots in generated guides; and remote troubleshooting, enabling support teams to request visual evidence without exposing sensitive windows. Because Screeny operates exclusively on macOS, it is a natural fit for Apple‑centric development environments that prioritize user privacy and background operation.

In summary, Screeny offers a secure, low‑overhead method for AI assistants to gain visual context on macOS. Its explicit approval system, background capture mode, and strict data handling make it a standout tool for developers who need reliable screenshot capabilities without compromising privacy.