DXT Weekly
Welcome to another issue of DXT Weekly, your guide to the latest Desktop Extensions transforming AI applications. This week brings a new (to us!) collection of DXTs that push the boundaries of what's possible when AI meets specialized platforms and services.
From enterprise-grade solutions like secure Heroku management and high-performance Trino database querying, to creative tools that connect Claude with LINE messaging, Blender 3D modeling, and academic research through arXiv LaTeX processing, this issue showcases the growing ecosystem of MCP-powered extensions. Whether you're automating trading strategies with Composer, organizing research with Raindrop.io, or exploring the cutting-edge metaprogramming capabilities of ClojureScript-based servers, these DXTs demonstrate how the Model Context Protocol continues to unlock new possibilities for AI-assisted workflows.
Connect Claude to LINE's 200 Million Users
by LY Corporation
This extension bridges Claude Desktop with LINE's Messaging API, letting you send messages, manage rich menus, and retrieve user profiles through LINE Official Accounts. With 435 GitHub stars, it's gaining traction among developers building automated customer service and marketing workflows. Good for businesses that want to deploy AI-powered chatbots on LINE, automate customer communications, or integrate Claude's capabilities into existing LINE-based workflows.
Manage Your Heroku Infrastructure Through Natural Language
by Heroku
The Heroku Platform MCP Server lets you control your entire Heroku infrastructure using plain English commands through Claude and other AI applications. Instead of switching between terminals and dashboards, you can deploy apps, scale resources, manage databases, and monitor logs directly through conversation with your AI assistant. It provides secure, authenticated access to the full range of Heroku platform capabilities.
Good for DevOps teams who want to streamline deployment workflows, developers managing multiple Heroku apps, or anyone who prefers conversational interfaces over command-line tools for infrastructure management.
Self-Modifying MCP Server with Metaprogramming
by MCP-PIF Team
This ClojureScript-based MCP server can modify its own code and capabilities while running, offering metaprogramming features that let it evolve and adapt its functionality dynamically. It includes memory storage, journaling, and the ability to execute and modify tools on the fly, making it a powerful foundation for developers who want their AI integrations to learn and improve over time.
Good for developers building adaptive AI workflows, prototyping self-improving systems, or creating MCP servers that need to evolve their capabilities based on usage patterns.
Connect Your AI Assistant to Trino's Distributed SQL Engine
by Tommy Nguyen
This MCP server bridges the gap between AI assistants and Trino's powerful distributed query engine, letting you execute SQL queries, explore database schemas, and analyze data structure directly through natural language conversations. Your AI can now discover catalogs and tables, retrieve column information, and run queries with built-in security restrictions and timeout controls.
Good for data analysts and engineers who want to explore large distributed datasets through conversational queries, or teams that need to give AI assistants secure access to their Trino infrastructure for automated reporting and data discovery.
Secure Business Data Access for AI Applications
by Entropy Data
This MCP server creates a secure bridge between your AI applications and business data systems. It provides controlled access to data products through search, retrieval, and query capabilities while maintaining proper access controls and security protocols. Teams can now let their AI assistants work with real business data without compromising security or bypassing governance frameworks.
Good for organizations that need to connect AI tools to internal databases, data warehouses, or analytics platforms while maintaining strict data governance and access controls.
Access Your Raindrop.io Bookmarks Directly in Claude
by Adam E
This DXT connects Claude to your Raindrop.io bookmark collection, letting you search and reference your saved articles, resources, and research materials without leaving your AI conversation. Instead of manually hunting through your bookmarks or switching between apps, you can have Claude pull relevant saved content directly into your workflow.
Good for researchers who need to reference saved articles during analysis, content creators building on previously bookmarked resources, or anyone who wants their curated bookmark collection available as context for AI conversations.
Turn Claude into Your Trading Strategy Partner
by Composer Technologies Inc.
The Composer DXT connects Claude to a full-featured trading platform where you can create, backtest, and automate investment strategies through natural conversation. Instead of manually coding trading algorithms or navigating complex interfaces, you can describe your investment ideas to Claude and have it build, test, and deploy strategies using real market data. The extension handles everything from portfolio monitoring to live trade execution across stocks and crypto.
Good for investors who want to test "what-if" scenarios before risking real money, or anyone looking to automate their investment approach without learning complex trading software.
Search Beyond Google: Comprehensive Search Through Brave's API
by Mike Chao
This extension brings Brave Search capabilities directly into Claude, letting you search the web, find images, browse news, discover videos, and locate points of interest without switching between applications. Instead of copying and pasting search results or juggling multiple browser tabs, you can gather information from multiple search types in a single conversation flow.
Good for research projects that need diverse content types, competitive analysis across different media formats, or when you want to cross-reference information from web results with current news coverage.
Process arXiv Papers with Full LaTeX Source Access
by Takashi Ishida
This DXT connects Claude to arXiv's LaTeX source files, giving you direct access to the raw mathematical expressions and formatting behind scientific papers. Instead of working with PDFs where complex equations might be garbled or misinterpreted, you get the original LaTeX code that authors used to write their papers.
Good for researchers analyzing mathematical proofs, students working through complex equations, or anyone who needs to understand the precise notation and structure of scientific papers without losing meaning in translation.
Control Blender 3D Models Through AI Conversation
by ahujasid
This DXT bridges Claude with Blender's 3D modeling environment, letting you manipulate scenes, objects, and animations through natural language commands. Instead of navigating complex menus and shortcuts, you can describe what you want to create or modify and have Claude execute the changes directly in your Blender workspace.
Good for rapid prototyping of 3D concepts, learning Blender workflows through conversational guidance, and automating repetitive modeling tasks without writing Python scripts manually.
What's Next
The DXT ecosystem is growing rapidly! We're tracking hundreds of extensions already, and counting.
Got a DXT to share? Drop us a line - we'd love to feature it in a future issue!
Want to build your own DXT? Check out Anthropic's DXT documentation: https://github.com/anthropics/dxt to get started.
DXT Weekly is an independent newsletter tracking the Desktop Extension ecosystem. Not affiliated with Anthropic.
Subscribe at https://dxtweekly.com | Suggest a DXT | Sponsor this newsletter