vibe-eyes
Vibe-Eyes is an MCP server designed to enhance interactions between large language models and browser-based games by providing real-time visual context and debug information. It captures and vectorizes canvas elements and exposes this data through the Model Context Protocol (MCP).
What is the primary purpose of Vibe-Eyes?
Vibe-Eyes is designed to provide LLMs with visual context from browser-based applications, enhancing their ability to assist in debugging by correlating visual states with console logs and errors.
How does Vibe-Eyes capture and process data?
Vibe-Eyes uses a lightweight browser client to capture canvas content and debug information, which is then sent to a Node.js server via WebSockets. The server vectorizes the images into SVG format and stores the debug data for LLM access.
Can Vibe-Eyes be integrated with any browser-based application?
Yes, Vibe-Eyes can be integrated with any browser-based application that uses canvas elements, by including the Vibe-Eyes client script and configuring it to communicate with the Vibe-Eyes server.
What kind of data does Vibe-Eyes provide to LLMs?
Vibe-Eyes provides LLMs with vectorized SVG representations of canvas content, console logs, errors, and unhandled exceptions, allowing LLMs to have a comprehensive view of the application's state.
Is Vibe-Eyes suitable for real-time applications?
Yes, Vibe-Eyes is designed to minimize performance impact and can be configured to capture data at intervals suitable for real-time applications, providing timely visual and debug information to LLMs.