Cline MCP Server for Bookmarks (VS Code Setup, 2026)

Cline is an autonomous AI coding agent that runs inside VS Code. It can read your codebase, edit files, run commands, and use MCP servers to connect to external data — including your ContextBolt bookmarks.

Connecting ContextBolt to Cline gives the AI agent access to your saved web pages and your saves from Twitter/X, Reddit and LinkedIn. During coding sessions, Cline can search your bookmarks to find relevant articles, discussions, and threads that inform the code it’s writing.

Why connect bookmarks to Cline

Cline works best when it has context. The more it knows about what you’re building and why, the better its output. Your bookmarks are a curated collection of technical knowledge you found worth saving — that’s exactly the kind of context that makes AI-generated code better.

When Cline is working on a task, it can search your bookmarks to find patterns and approaches you’ve already vetted. If you saved a detailed Reddit discussion about handling race conditions in async code, Cline can reference that when writing concurrent code for your project.

The auto-approve feature makes this especially smooth. Once you trust the ContextBolt search tools, add them to Cline’s auto-approve list and it will search your bookmarks freely as part of its workflow, without asking permission each time.

Cline MCP configuration file

Use https://api.contextbolt.app/mcp and complete the browser sign-in if your client supports remote MCP with OAuth. Enter your Pro account email and email code, then allow the connection. For a client that cannot complete that flow, use the personal-URL compatibility configuration shown here. Copy the full private URL from the extension under Settings, MCP Integration, Older clients and replace the example URL with it.

Cline stores MCP config at:

• Windows: %APPDATA%\Code\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json
• macOS: ~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
• Linux: ~/.config/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json

Add this to the file, replacing the whole https://api.contextbolt.app/mcp/YOUR_TOKEN example with your private URL from the ContextBolt extension under Settings → MCP Integration. Newer versions keep it under Older clients.

{
  "mcpServers": {
    "bookmarks": {
      "url": "https://api.contextbolt.app/mcp/YOUR_TOKEN",
      "disabled": false,
      "autoApprove": []
    }
  }
}

When we tested this on Cline 3.x in VS Code, the panel picks up the new server within a second of saving the file. To enable auto-approve, add the read tool names (search_bookmarks, list_clusters, get_cluster_bookmarks, get_recent_bookmarks) to the autoApprove array. Cline will then call them without prompting, which keeps long agentic sessions uninterrupted. The fifth tool, save_bookmark, writes a new bookmark, so leave it off the auto-approve list if you would rather confirm each save.

If you are on an older Cline release that does not yet read "url" for remote servers, the FAQ at the bottom of this page covers the mcp-remote bridge as a workaround.

Tips for Cline workflows

Enable auto-approve for bookmark search: After verifying the tools work correctly, add them to Cline’s auto-approve list. This lets Cline search your bookmarks as part of its autonomous workflow without interrupting for permission.

Front-load context: At the start of a Cline session, ask it to search your bookmarks for the topic you’re working on. This gives it relevant material before it starts writing code.

Combine with file reads: Ask Cline “Check my bookmarks for content about [pattern], then review how it’s implemented in [file].” This connects your external knowledge to your codebase.