Jervis Desktop MCP

Connect AI assistants to local project folders through MCP, with access permissions, device pairing and pause controls.

  • Project Type : Desktop platform · MCP integration
  • Share :

Discuss your project

Share your requirements and the parts of your workflow you want to improve.

Jervis Desktop MCP

Project Links

Project Overview

Jervis Desktop MCP connects AI assistants to project folders on your computer. You choose which folders to share, set read-only or editing access, pair the device and review activity. Sharing can be paused from the desktop app.

The same MCP runtime serves local clients through stdio and remote clients through Streamable HTTP. File operations stay within the selected workspace, and editing requires explicit permission. The Windows companion is available as a preview in Arabic and English.

The Challenge

  • Give assistants access to the right files within defined workspace boundaries.
  • Use the same tools and permission rules for local and remote clients.
  • Make device pairing, activity and access controls clear to the user.

How we built it

A TypeScript runtime checks permissions before executing MCP tools. The Electron companion manages folders, pairing and sharing, with read-only access as the default. Local and remote connections use the same tool definitions.

Technology Stack

  • TypeScript
  • MCP
  • Electron
  • Node.js

What the product Delivers

AI assistants can work with selected local projects while the user controls folder access, editing permissions and when sharing is active.