The Problem: Information Overload
As developers, we're constantly saving links. And that's not a real problem. The problem arises when we are trying to recall that blog post that covers "How to configure MCP with Streamable HTTP inside of a Lambda function". Or if we are trying to list all of our bookmarks related to a specific topic. For me this has always been nothing but frustration.
The Solution
To solve this problem I built LinkVault, which is an open-source MCP server that allows you to integrate with AI Agents like Amazon Q CLI to provide AI-powered bookmark management. So, instead of building another standalone tool, you can use Natural Language in a conversational interface to manage your bookmarks.
Key Features
- Amazon Q CLI Integration: Natural language bookmark management right in your terminal
- MCP Server Architecture: Extends Amazon Q with custom bookmark management tools
- Chrome Multi-Profile Support: Access bookmarks from all your Chrome profiles
- Content Extraction: Automatically analyze and categorize webpage content
- Cross-Platform: Works wherever Amazon Q CLI runs (macOS, Windows, Linux)
The Power of Model Context Protocol
The real innovation in LinkVault is how it uses the Model Context Protocol (MCP) to extend Amazon Q's capabilities. MCP allows Amazon Q to access external tools and data sources, effectively teaching it new skills.
When you register LinkVault as an MCP server with Amazon Q CLI, it gains the ability to:
- Extract and analyze content from web URLs
- Access your Chrome bookmarks across all profiles
- Store, categorize, and tag bookmarks intelligently
- Search across your entire bookmark collection using natural language
Amazon Q CLI in Action
Here's what the experience looks like in your terminal:
Checking if a bookmark exists.
$ q chat
Welcome to Amazon Q CLI!
> > Can you tell me if I have this bookmark stored? https://community.aws/content/2vzj07Wyk6Lw281Tvs1Lw7kJJNW/building-scalable-mcp-servers-on-aws-lambda-a-practical-guide?lang=en
I'll check if you have that bookmark stored. Let me search for it in your bookmarks.
🛠️ Using tool: search_bookmarks from mcp server bookmark_manager
⋮
● Running search_bookmarks with the param:
⋮ {
⋮ "arguments": {
⋮ "query": "https://community.aws/content/2vzj07Wyk6Lw281Tvs1Lw7kJJNW/building-scalable-mcp-servers-on-aws-lambda-a-practical-guide"
⋮ },
⋮ "name": "search_bookmarks"
⋮ }
Allow this action? Use 't' to trust (always allow) this tool for the session. [y/n/t]:
> t
⋮
● Completed in 0.3s
I've searched your bookmarks and couldn't find this URL:
https://community.aws/content/2vzj07Wyk6Lw281Tvs1Lw7kJJNW/building-scalable-mcp-servers-on-aws-
practical-guide
The search returned 0 results, which means this bookmark is not currently stored in your collection. Would
you like me to add it to your bookmarks?
Adding a new bookmark
> yes add it
I'll add that bookmark for you. Let me first get some information about the URL to create a proper bookmark
entry.
🛠️ Using tool: get_url_data from mcp server bookmark_manager
⋮
● Running get_url_data with the param:
⋮ {
⋮ "arguments": {
⋮ "url": "https://community.aws/content/2vzj07Wyk6Lw281Tvs1Lw7kJJNW/building-scalable-mcp-servers-on-aws-lambda-a-practical-guide"
⋮ },
⋮ "name": "get_url_data"
⋮ }
Allow this action? Use 't' to trust (always allow) this tool for the session. [y/n/t]:
> t
⋮
● Completed in 0.384s
🛠️ Using tool: store_url from mcp server bookmark_manager
⋮
● Running store_url with the param:
⋮ {
⋮ "arguments": {
⋮ "category": "AWS",
⋮ "description": "A practical guide on building scalable Model Context Protocol (MCP) servers using AWS Lambda and Lambda Web Adapter, covering implementation of sessionless Streamable HTTP communication, Lambda Function URLs, and performance optimization.",
⋮ "importance": 4,
⋮ "notes": "Useful resource for implementing MCP servers in a serverless architecture with AWS Lambda. Contains implementation steps, optimization tips, and best practices.",
⋮ "tags": [
⋮ "MCP",
⋮ "Lambda",
⋮ "Serverless",
⋮ "AWS",
⋮ "Model Context Protocol"
⋮ ],
⋮ "title": "Building Scalable MCP Servers on AWS Lambda: A Practical Guide",
⋮ "url": "https://community.aws/content/2vzj07Wyk6Lw281Tvs1Lw7kJJNW/building-scalable-mcp-servers-on-aws-lambda-a-practical-guide"
⋮ },
⋮ "name": "store_url"
⋮ }
Allow this action? Use 't' to trust (always allow) this tool for the session. [y/n/t]:
> t
⋮
● Completed in 0.3s
I've successfully added the bookmark to your collection:
• Title: Building Scalable MCP Servers on AWS Lambda: A Practical Guide
• URL: https://community.aws/content/2vzj07Wyk6Lw281Tvs1Lw7kJJNW/building-scalable-mcp-servers-on-aws-
lambda-a-practical-guide
• Category: AWS
• Tags: MCP, Lambda, Serverless, AWS, Model Context Protocol
• Importance: 4/5
• Date Added: 2025-05-28
The bookmark has been stored with a description that summarizes the article's content about implementing
MCP servers using AWS Lambda and Lambda Web Adapter. I've also added notes highlighting that it's a useful
resource for serverless MCP implementation with implementation steps and optimization tips.
This natural language interface makes bookmark management feel effortless for me since its right in my Amazon Q Cli daily workflow. Yes, I use it for everything! If you have questions on how I use it let me know in the comments and I will write a new post.
How MCP Makes It Possible
The Model Context Protocol is what enables this seamless integration. LinkVault exposes a set of tools to Amazon Q:
Multi-Profile Chrome Integration
One feature that I am working to improve is the ability for LinkVault to access bookmarks across all your Chrome profiles. If you're like me, you might have separate profiles for:
- Work (AWS documentation, business IP, customer architectures)
- Personal / Home (tech blogs, learning resources)
- Development (code samples, GitHub repositories)
I've designed LinkVault to bring them all together in one unified interface.
NOTE: This feature will be improved to support Firefox and Safari soon.
Setting Up LinkVault with Amazon Q CLI
Getting started with LinkVault is straightforward:
- Install Amazon Q CLI if you haven't already
- Clone the LinkVault repository
git clone https://github.com/labeveryday/linkvault-mcp-server.git && cd linkvault-mcp-server
- Install UV to manage your environment setup
- Create a virtual environment
uv init
uv venv
source .venv/bin/activate
- Install dependencies with
uv install . - Register LinkVault as an MCP server in your Amazon Q configuration:
// ~/.aws/amazonq/mcp.json
{
"bookmark_manager": {
"command": "uv",
"args": ["--directory", "/path/to/linkvault-mcp-server", "run", "src/server.py"],
"env": {},
"disabled": false,
"autoApprove": ["get_url_data", "store_url", "search_bookmarks", "list_chrome_bookmarks"]
}
}
- Start a new Amazon Q CLI session and begin managing your bookmarks!
Building Your Own MCP Tools
The beauty of the Model Context Protocol is that it allows developers to extend Amazon Q CLI's and AI agent capabilities in countless ways. LinkVault is just one example of what's possible.
By following a similar pattern, you could build MCP servers that:
- Connect Amazon Q CLI to your internal knowledge bases
- Provide custom code generation capabilities
- Enable interaction with specialized APIs
- Create AI-powered interfaces to your development tools
What's Next for LinkVault?
I'm actively working on expanding LinkVault with:
- Safari and Firefox browser support
- Enhanced content analysis for better categorization
- Integration with AWS service documentation
- Collaborative bookmark sharing for teams
- Host in AWS
Conclusion
LinkVault demonstrates the power of extending Amazon Q CLI through the Model Context Protocol. By connecting Amazon Q CLI to your browser bookmarks, it creates a natural language interface to your web research that fits seamlessly into your development workflow.
If you're an AWS developer looking to tame your bookmark chaos while staying in your terminal, give LinkVault a try. And if you're interested in building your own MCP tools, I hope this project provides some inspiration.
Have you built tools that integrate with Amazon Q CLI? I'd love to hear about your experiences in the comments!
This post was written by Du'An Lightfoot, Sr. Developer Advocate at AWS. The views and opinions expressed in this post are my own and don't necessarily represent the positions, strategies, or opinions of Amazon Web Services.
Top comments (0)