Cursor MCP Server for Bookmarks (2026 Setup, 3 min)

Cursor is an AI-powered code editor built on VS Code. It supports MCP (Model Context Protocol), which means you can give its AI features direct access to your ContextBolt bookmarks while you code.

Once connected, Cursor’s AI chat and Composer can search your saved content from Twitter/X, Reddit, and LinkedIn. This is useful when you’re building something and want to reference articles, discussions, or technical threads you bookmarked weeks or months ago.

Why this integration matters for Cursor users

Cursor users already rely on AI to help with code. The ContextBolt integration extends that AI with your own curated knowledge.

Think about how you work. You read a great thread on Twitter about error handling patterns, you bookmark it. You save a Reddit post explaining database indexing strategies. Three weeks later, you’re building exactly the feature those saves would help with, but they’re buried in your browser.

With this integration, you can ask Cursor’s AI to search your bookmarks mid-conversation. The search is semantic, so you describe what you need in natural language. Cursor already understands your codebase — now it can also understand what you’ve been reading and saving.

The full Cursor MCP config block

Drop this into .cursor/mcp.json (project) or ~/.cursor/mcp.json (global), with YOUR_TOKEN swapped for the token from your ContextBolt extension settings:

{
  "mcpServers": {
    "bookmarks": {
      "url": "https://api.contextbolt.app/mcp/YOUR_TOKEN"
    }
  }
}

When we tested this on Cursor 0.48 and 0.49, the connection picks up on window reload without restart. If the bookmarks server stays gray, the most common cause is a Cursor version below 0.48: pin the version in Cursor’s release notes, or fall back to the mcp-remote bridge in the FAQ.

Tips for Cursor workflows

Use in Composer for context-rich generation. When starting a Composer session, ask Cursor to check your bookmarks for relevant patterns first. This grounds the generated code in material you’ve already vetted.

Reference saves in code reviews. If you saved an article about a pattern you’re now seeing in a PR, ask Cursor to find it and summarize the relevant points.

Combine with codebase search. Ask Cursor something like “Search my bookmarks for WebSocket best practices, then review how our current WebSocket implementation compares.” This connects your saved knowledge to your actual code.

How to connect ContextBolt to Cursor

1. Get your MCP token from ContextBolt

   Open the ContextBolt extension popup and click Settings. Under the MCP section, click 'Copy MCP Token'. MCP is a Pro feature ($6/month) and cloud sync must be enabled in the same panel, otherwise the server returns no data.

2. Open Cursor's MCP config

   Create or edit .cursor/mcp.json in your project root, or the global file at ~/.cursor/mcp.json. Remote HTTP server support requires Cursor 0.48.0 or newer; older Cursor versions need the mcp-remote bridge (covered in the FAQ below).

3. Add the ContextBolt MCP server

   Paste this under mcpServers and replace YOUR_TOKEN with the token from step one: {"bookmarks": {"url": "https://api.contextbolt.app/mcp/YOUR_TOKEN"}}. Cursor uses "url" directly with no "type" field, which is the most common mistake people copy across from Claude Code config.

4. Verify the connection

   Save the file and reload the Cursor window. Open Settings, find the MCP Servers section, and you should see bookmarks listed with a green status. Start an AI chat and ask 'Search my bookmarks for API design patterns' to confirm tools are firing.

Example prompts for ContextBolt + Cursor

Once connected, try asking Cursor:

What you can do with ContextBolt in Cursor