Using MCP

Connect your AI assistant to Contextium so it is grounded in your team's shared knowledge — files, skills, agents, and workflows. Instead of each person copy-pasting different context into their chat, every assistant pulls from the same source of truth.

Works with: Claude Desktop, VS Code Copilot, and Cursor. Any tool that supports remote MCP connections will work.

---

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external tools. When you add Contextium's MCP connection, your AI can browse your libraries, search your files, load skills and agents, run workflows, and create or edit content — all from the chat interface. This is how you ground your assistant in real, up-to-date context instead of relying on the model's training data.

---

Connect your tool

Contextium runs a remote MCP server, so there is nothing to install on your computer. You just paste a short snippet into your AI tool's settings.

Claude Desktop

1. Open Claude Desktop and go to Settings
2. Select Connectors
3. Click Add Custom Connector
4. Enter the URL: https://mcp.contextium.io
5. Click Add

VS Code

To add the Contextium remote MCP server to VS Code:

1. Press CTRL+SHIFT+P and search Add Server, select MCP: Add Server.
2. Select HTTP
3. Enter the URL: https://mcp.contextium.io and press Enter
4. Enter the name Contextium and press Enter
5. Press Enter to select to save to the Global workspace.

The resulting MCP server configuration can be found in mcp.json as follows:

{
  "servers": {
    "contextium": {
      "url": "https://mcp.contextium.io"
    }
  }
}

Cursor

Go to Settings > MCP and add:

{
  "mcpServers": {
    "contextium": {
      "type": "http",
      "url": "https://mcp.contextium.io"
    }
  }
}

---

Sign in

The first time you use a Contextium tool, your AI tool opens a browser window so you can sign in with your Contextium account. Once you've signed in, the session stays active — you won't need to sign in again unless you log out.

---

Check it's working

Restart your AI tool, then type:

"List my Contextium workspaces"

If you see your workspaces listed, you're connected.

---

What you can do

Browse your workspace

Start broad, then drill down:

• "List my workspaces" — see everything you have access to
• "Show libraries in [workspace]" — see your context libraries with file counts
• "List files in [library]" — see file names and details
• "Show me [filename]" — read the full content of a specific file

Search your documentation

• "Search my docs for authentication" — find content across your whole workspace
• "Find files about onboarding" — look up files by name
• "Find all files tagged #security" — filter by tag

Create and update content

• "Create a library called API Docs" — add a new context library
• "Create a file called auth-guide.md in API Docs" — add a new file
• "Update auth-guide.md with this content: ..." — edit an existing file; Contextium saves a version history so you can always go back

---

Troubleshooting

"Authentication failed" or you're stuck in a sign-in loop: Make sure you're signing in with the account that has access to the workspace. Try clearing your browser session and signing in again.

"No workspaces found": Check that your account is a member of at least one workspace at contextium.io/dashboard.

---

For the full list of tools and their parameters, see the MCP Tool Reference.