Connect Any Agent to Business Central
Connect any agent to Business Central, even on Mac or Linux. I built a cross-platform MCP proxy that makes it very simple — one command to install and configure.
MCPs for AL Vibe Coding
The AL Vibe Coding initiative started as a simple idea to make copilots understand and write great AL code — and quickly grew into a movement powered by MVPs and Microsoft engineers. This post explores the most useful MCP servers born from that collaboration, with stories, tools, and install links for AL developers.
Microsoft Learn MCP Is Live!
Now your VSCode GitHub Copilot or Cursor can answer questions using Microsoft Learn. Install in 1-click.
CentralQ Data Insights Agent is Live!
Imagine asking Business Central about your data... and it just answers you — in plain text, with charts and all. Today, this is real.
Code Interpreter for Business Central
Unlock AI-powered data analysis in Business Central with the open-source Code Interpreter, combining LLMs, Python, and real-time insights.
CentralQ: What We’ve Achieved Together
CentralQ reflects on 1.5 years of growth, answering 300,000 questions with the power of community knowledge and Microsoft Learn.
Directions 2024: Where to Find Me
Catch me at Directions EMEA 2024 in Vienna, where I'll be diving into AI's future in Business Central with inspiring sessions and incredible co-speakers—don’t miss it!
Tools for Agents
Discover the latest advancements in the Copilot Toolkit with Business Central 25.0, including auto tools invocation and parallel function calling, paving the way for more efficient autonomous agents.
Number Series Copilot
Explore my open-source contribution, the Number Series Copilot, which simplifies number series management in Business Central and saves user's valuable time.
Copilot Toolkit: Function Calling
Discover how to enhance your Business Central capabilities with function calling in the AL Copilot Toolkit. Learn to seamlessly integrate natural language queries with internal business logic, providing intuitive and powerful responses without exposing sensitive data.
