StayPut

What StayPut MCP server does

StayPut MCP server is intended for local document and photo processing. The supplied description identifies PDF operations for merging files, splitting or extracting pages, rotating pages, and adding page numbers. It also identifies photo processing for stripping GPS metadata. These operations are useful when files contain sensitive content or location information that should not leave the device.

The broader StayPut project is presented as a browser-based collection of tools for images, PDFs, video, and audio. Its pages are built from static HTML, JavaScript, and WebAssembly, with processing performed in the browser rather than on a backend. The MCP listing specifically highlights the local PDF and photo capabilities, so this entry does not treat every browser tool as an MCP capability.

How it works

StayPut MCP server is described as local-first: the files remain on the user’s device while the selected operation runs locally. The project documentation says that files are not uploaded and that there is no backend receiving them. After a page has been visited, the web tools can continue working offline through browser caching; AI-model tools are an exception because their model must first be downloaded during an online run.

This design suits workflows where an agent needs to prepare or clean files without sending their contents to a hosted conversion service. It also means the device and its available storage determine what can be processed. The provided material does not specify how an MCP client passes files to the server or whether processing is implemented through a browser session, a local executable, or another transport.

Setup and configuration

The available material does not include MCP installation commands, transport settings, command-line options, configuration files, or environment variables. It identifies the public StayPut site and repository, but those references alone are not enough to document a reproducible MCP setup. Operators should consult the repository’s MCP-specific instructions before configuring a client.

For the browser-based StayPut tools, the documented workflow is to open a tool, add a file, and process it in the tab. The project also describes an offline-after-first-visit mode for its web pages. No MCP client names or client-specific configuration are provided here.

Tools and capabilities

The documented MCP-focused capabilities are:

• Merge PDF files locally.
• Split PDFs or extract pages.
• Rotate PDF pages.
• Add page numbers to PDFs.
• Remove GPS metadata from photos.

The project’s broader tool list also mentions EXIF removal and viewing, but the supplied MCP description specifically calls out stripping photo GPS data. StayPut MCP server should therefore be evaluated against the narrower PDF and photo workflow rather than assumed to expose every media feature in the website.

Limitations and notes

StayPut MCP server is not documented here as a general-purpose cloud file service. Its stated local-processing model avoids uploads, but large files still depend on the device’s available resources. The material does not provide performance limits, supported operating systems, file-size guarantees, error behavior, or details about preserving metadata beyond the GPS-removal use case.

The README labels the project MIT-licensed and says that the web application has no server-side backend. It does not name compatible MCP clients. Client support, installation method, and the exact MCP tool schemas remain unspecified in the supplied excerpt.