News

VS Code Goes Transparent as Open-Source AI Editor

• By David Ramel
• 06/30/2025

Microsoft today announced the first milestone of its project to transform the popular Visual Studio Code editor into an open-source AI development environment. Specifically, the company revealed the GitHub Copilot Chat extension is now open source on GitHub under the MIT license.

Transparency, Trust and Open Development
According to Microsoft, the decision to open source GitHub Copilot Chat stems from a growing demand for transparency in how AI-assisted developer tools work -- particularly around prompt engineering, data collection, and model interaction.

"We believe AI experiences can thrive by leveraging the vibrant open-source community -- just as VS Code has successfully done over the past decade," the VS Code team wrote in its blog post announcing the milestone. "As AI is becoming an integral part of the modern coding experience, it should be developed openly alongside VS Code itself."

The popular VS Code chat tool has garnered more than 35 million installs at the time of this writing.

Microsoft listed the reasons for the open sourcing specifically:

• Large language models have significantly improved, mitigating the need for "secret sauce" prompting strategies.
• The most popular and effective UX treatments for AI interactions are now common across editors. We want to enable the community to refine and build on these common UI elements by making them available in a stable, open codebase.
• An ecosystem of open source AI tools and VS Code extensions has emerged. We want to make it easier for these extension authors to build, debug, and test their extensions. This is especially challenging today without access to the source code in the Copilot Chat extension.
• We've gotten a lot of questions about the data that is collected by AI editors. Open sourcing the Copilot Chat extension enables you to see the data we collect, increasing transparency.
• Malicious actors are increasingly targeting AI developer tools. Throughout VS Code's history as OSS, community issues and PRs have helped us find and fix security issues quickly.

As implied, the move comes amid long-running concerns over telemetry and data privacy in Microsoft's development tooling. While VS Code itself is built on the open source Code - OSS base, the official VS Code distribution includes proprietary Microsoft elements such as branding, telemetry, and integrations with closed backend services. By contrast, the GitHub Copilot Chat extension -- now licensed under MIT -- offers a look under the hood at what information is sent to language models and how responses are generated, structured, and applied to users' codebases.

Microsoft also points to a broader shift in the developer ecosystem. As large language models (LLMs) improve and design patterns for AI coding assistants mature, the company sees less need for "secret sauce" implementation tactics and more value in fostering a community that can co-develop the next generation of AI-enhanced tooling.

Inside the Project: What's Open, What's Not
The open-sourced GitHub Copilot Chat extension includes full access to agent mode logic, system prompts, and even the telemetry mechanisms used to track usage -- all hosted in a public GitHub repository under active development by the VS Code team and community contributors. Microsoft says the next step is to begin refactoring parts of the extension into VS Code core itself, bringing AI more deeply into the editor's foundational architecture.

However, the core Copilot services -- such as model serving infrastructure and the Copilot Completions extension that provides inline suggestions -- remain closed source. In their place, the open Copilot Chat extension will eventually offer equivalent capabilities, enabling a more unified and open experience for chat-based and inline code assistance.

Developers can already browse the codebase, submit pull requests, and file issues. Microsoft has also pledged to open source its prompt test infrastructure, which will help ensure community PRs remain stable and testable despite the non-deterministic nature of LLM responses.

What It Means for Developers
For VS Code extension authors and enterprise developers, the open sourcing of Copilot Chat is more than a transparency gesture -- it's a practical shift that could unlock new workflows and deeper integrations. Extension developers, for example, now have visibility into how Microsoft handles prompt context, telemetry, and multi-step agent interactions. That knowledge can inform the design of third-party AI extensions and help avoid duplication of effort.

For security-conscious organizations, access to the source code could provide critical insight into what's collected during AI-powered sessions -- a recurring point of concern for teams bound by compliance and governance policies. "Open sourcing the Copilot Chat extension enables you to see the data we collect," Microsoft said in announcing the initiative last month, framing the move as a step toward demystifying the black-box nature of AI tooling.

What's Next
With the GitHub Copilot Chat extension now public, Microsoft's next focus is refactoring key components into the VS Code core. This deeper integration aims to make AI features more native to the editor, rather than functioning solely as a separate extension. The team has also committed to open sourcing its prompt testing infrastructure to support reliable community contributions, despite the inherent variability in LLM responses. Long term, the goal is to enable AI-enhanced development workflows that are open by default -- both in visibility and extensibility.

While the backend remains proprietary, Microsoft's goal seems to be to create an AI-enabled development environment where the client-side logic is community-auditable, testable, and adaptable -- without sacrificing user experience or developer productivity. In the coming months, the community will be watching to see just how much of that vision is realized. More information is available in a FAQ.