MCP Servers

Model Context Protocol (MCP) servers for use with Claude Code and other MCP-compatible clients.

Servers

@10xdeca/mcp-kan

MCP server for Kan board task management. Provides tools for managing workspaces, boards, lists, cards, labels, checklists, and comments.

Environment variables:

• KAN_API_KEY (required) — your Kan API key
• KAN_BASE_URL (optional) — defaults to https://kan.bn/api/v1

@10xdeca/mcp-outline

MCP server for Outline wiki. Provides tools for managing documents, collections, comments, users, and revisions.

Environment variables:

• OUTLINE_API_KEY (required) — your Outline API key
• OUTLINE_BASE_URL (optional) — defaults to https://kb.xdeca.com/api

@10xdeca/mcp-radicale

MCP server for Radicale CalDAV/CardDAV. Provides tools for managing calendars, events, todos, address books, and contacts.

Environment variables:

• RADICALE_URL (required) — your Radicale server URL (e.g. https://dav.xdeca.com)
• RADICALE_USERNAME (required) — Radicale username
• RADICALE_PASSWORD (required) — Radicale password

Setup

npm install

Usage with Claude Code

Add to your ~/.claude/claude_desktop_config.json or MCP settings:

{
  "mcpServers": {
    "kan": {
      "command": "node",
      "args": ["./packages/kan/index.js"],
      "env": {
        "KAN_API_KEY": "your-api-key"
      }
    },
    "outline": {
      "command": "node",
      "args": ["./packages/outline/index.js"],
      "env": {
        "OUTLINE_API_KEY": "your-api-key"
      }
    },
    "radicale": {
      "command": "node",
      "args": ["./packages/radicale/index.js"],
      "env": {
        "RADICALE_URL": "https://your-radicale-server.com",
        "RADICALE_USERNAME": "your-username",
        "RADICALE_PASSWORD": "your-password"
      }
    }
  }
}

---

---

Using This MCP Server with Jan AI

Author: Prompted by SophieA; Generated by Claude Sonnet 4.6;

This guide assumes:

• You already have Jan AI installed → https://jan.ai
• You are using a model that supports MCP (wrench icon)
• You have received an OUTLINE_API_KEY from your admin

Note: This guide was written using macOS. Paths, commands, and environment variable setup may differ slightly on Linux. Adjust accordingly.

1️⃣ Clone the Repository

git clone https://github.com/10xdeca/mcp-servers
cd mcp-servers

2️⃣ Install Dependencies

From the project root:

npm install

This installs dependencies for both Kan and Outline MCP servers.

3️⃣ Verify the Outline MCP Server Works

Run it manually once:

OUTLINE_API_KEY=your_key node packages/outline/index.js

If:

• It does not throw an error
• It just sits there running

That means it started correctly.

Press Ctrl+C to stop it.

If there are no errors, Jan AI will be able to connect to it.

4️⃣ Add MCP Server in Jan AI

Open Jan AI.

1. Go to Settings
2. Go to MCP Servers
3. Click Add MCP Server

Fill the Modal Like This

• Server Name: Outline xdeca
• Transport Type: STDIO
• Command: node
• Arguments:

/Users/taro/projects/mcp/sprints-xdeca-mcp/packages/outline/index.js

⚠️ Must be the full absolute path (no ~).

• Environment Variables:

OUTLINE_API_KEY = your_key_from_admin

Then toggle the server ON.

Alternative: Add via JSON

At the top right of the Add MCP Server modal, click:

Add server by JSON

Paste:

{
  "Outline xdeca": {
    "active": true,
    "args": [
      "/Users/you-user/mcp-servers/packages/outline/index.js"
    ],
    "command": "node",
    "env": {
      "OUTLINE_API_KEY": "your_key_here"
    },
    "type": "stdio"
  }
}

Save.

Sophie note:

At this point the Outline MCP server should toggle green. If it does not toggle green, on JanAI go to "Settings > General > Open Logs" to start debugging.

If it does toggle green, you can use it as is or you can create a new assistant for better results:

5️⃣ Create a Dedicated Outline Assistant

To make usage cleaner, create a new assistant in Jan:

1. Go to Settings → Assistants
2. Click Create Assistant
3. Name it: Outline Assistant
4. Select your MCP-capable model (must show the 🔧 wrench icon)
5. Paste your instructions from:

janai-mcp-outline-prompt.md

into the System Instructions field.

Save.

6️⃣ Use It

Now you can ask things like:

• “Search for onboarding docs”
• “Create a new document in Engineering collection”
• “Update the sprint planning document”
• Upload an image and ask it to create notes

The assistant will automatically call the Outline MCP tools.

---

That’s it. You now have a fully local Outline MCP integration running inside Jan AI.