📋 RECOMMENDED — Useful improvements. Plan your upgrade.

Version: 1.119 | Released: 2026-05-06 | Upgrade from: 1.118

Release at a Glance

Visual Studio Code 1.119 lands with significant enhancements for AI-assisted development, particularly focusing on agent capabilities and observability. This minor release, dated May 6, 2026, is a strong recommendation for developers leveraging VS Code’s AI features.

Here’s the TL;DR for developers:

  • Agent Browser Sharing empowers AI agents to directly interact with web pages within VS Code, streamlining web development workflows by allowing agents to edit, reload, and confirm fixes in a single turn.
  • OpenTelemetry Tracing introduces enhanced observability for AI coding sessions, providing deeper insights into agent activities and performance for better debugging and understanding.
  • Plan Agent Improvements further refine the Plan agent’s ability to tackle complex tasks, leading to more effective planning and execution.
  • Mermaid Diagrams in Chat now render directly within chat responses, enhancing visual communication and documentation within conversational AI sessions.

This release primarily targets developers engaged in AI-driven coding, especially those in web development or requiring detailed insights into agent behavior.

Headline New Features

VS Code 1.119 doubles down on its commitment to agent-first development, introducing features that make AI agents more interactive and transparent.

Agent Browser Sharing: Interactive Web Development

What it is: Agent Browser Sharing allows AI agents within VS Code to directly interact with web pages. This isn’t just about reading content; agents can now perform actions like editing code, triggering page reloads, and verifying fixes—all within the integrated environment.

Why it matters: For web developers, this feature dramatically shortens the feedback loop. Traditionally, an agent might suggest a code change, and the developer would then manually apply it, switch to a browser, refresh, and confirm. With Agent Browser Sharing, the agent can perform these steps autonomously.

Real-world insight: Imagine an agent tasked with fixing a UI bug. It can now:

  1. Identify the problematic HTML/CSS.
  2. Suggest and apply a code change in your workspace.
  3. Open an embedded browser, navigate to the page, and reload it.
  4. Visually confirm the fix (or identify if further changes are needed).

This integrated workflow reduces context switching and accelerates the iterative process of web development, pushing us closer to truly autonomous coding assistants.

OpenTelemetry Tracing for AI Sessions

What it is: This release integrates OpenTelemetry tracing into AI coding sessions. OpenTelemetry is a collection of tools, APIs, and SDKs used to instrument, generate, collect, and export telemetry data (metrics, logs, and traces) to help you analyze your software’s performance and behavior. In VS Code, it now specifically tracks agent activities.

Why it matters: As AI agents become more sophisticated, understanding their decision-making process and execution flow becomes critical. When an agent produces unexpected results or gets stuck, traditional debugging methods often fall short.

Core concept: OpenTelemetry tracing provides a detailed, timeline-based view of an agent’s internal operations. Each step an agent takes—from parsing a prompt, calling a tool, generating code, to interacting with the environment—can be represented as a “span” in a trace.

Breakdown:

flowchart TD DEV[Developer Prompt] --> AGENT_START[Agent Session Start] AGENT_START --> PLAN[Plan Task] PLAN --> TOOL_CALL[Call Tool API] TOOL_CALL --> CODE_GEN[Generate Code] CODE_GEN --> BROWSER_INTERACT[Browser Interaction] BROWSER_INTERACT --> CONFIRM[Confirm Fix Output] CONFIRM --> AGENT_END[Agent Session End] subgraph Trace["OpenTelemetry Trace"] AGENT_START -->|Span| PLAN PLAN -->|Span| TOOL_CALL TOOL_CALL -->|Span| CODE_GEN CODE_GEN -->|Span| BROWSER_INTERACT BROWSER_INTERACT -->|Span| CONFIRM end

This tracing allows developers to:

  • Pinpoint bottlenecks: Identify where agents spend most of their time.
  • Debug agent logic: Understand the sequence of operations that led to a particular outcome.
  • Optimize agent performance: Gain insights to refine agent prompts or tool usage.

Real-world insight: If an agent consistently fails to resolve a specific type of bug, OpenTelemetry traces can reveal if it’s struggling with prompt interpretation, tool selection, or execution errors, enabling more targeted improvements.

Plan Agent Improvements

The Plan agent, responsible for breaking down complex tasks into actionable steps, receives further enhancements. These improvements lead to more robust task planning and execution, making the agent more reliable for multi-step operations. This directly contributes to the agent’s ability to handle more sophisticated requests and deliver more accurate, structured outputs.

Mermaid Diagrams in Chat Responses

Visual communication is key. VS Code’s chat responses now support rendering Mermaid diagrams directly. This means when an agent or another user shares a system architecture, a flowchart, or a sequence diagram in chat, it will render visually rather than as raw text. This significantly improves clarity when discussing complex technical concepts or system designs.

Inline Chat UI Improvements

Various refinements have been made to the inline chat user interface. These improvements aim for a smoother, less intrusive, and more intuitive conversational AI experience directly within the code editor. This includes better context handling, improved response formatting, and general usability enhancements.

Developer Workflow Impact

This release profoundly impacts developers who are integrating AI agents into their daily coding practices, particularly in web development and those seeking deeper insights into AI behavior.

Accelerated Web Development with Agent Interaction

The most immediate impact of Agent Browser Sharing is on web development workflows. Developers can expect:

  • Faster Iteration Cycles: Agents can now perform a complete “fix-and-verify” loop without manual intervention, reducing the time spent on repetitive tasks like refreshing browsers.
  • Reduced Context Switching: Staying within the VS Code environment for both code editing and browser interaction streamlines focus and improves productivity.
  • Enhanced Agent Autonomy: Agents can tackle more complex web-related tasks end-to-end, from identifying issues to confirming visual changes.

This capability pushes the boundary of what AI agents can achieve in front-end development, making them more active participants in the development process rather than just code generators.

Deeper Observability for AI-Assisted Coding

OpenTelemetry Tracing is a game-changer for anyone building with or relying heavily on AI agents.

  • Debugging AI Behavior: Developers can now “look inside” the agent’s thought process. This is invaluable when an agent’s output is incorrect or unexpected, allowing for precise identification of where the logic went awry.
  • Performance Tuning: By analyzing traces, developers can identify inefficient agent prompts, slow tool calls, or redundant steps, leading to more optimized and faster AI assistance.
  • Trust and Transparency: Understanding why an agent made a certain decision builds trust in the AI system and helps developers learn how to better prompt and utilize their agents.

This feature moves AI agent usage from a black-box experience to a transparent, debuggable workflow, which is crucial for professional development environments.

Streamlined Communication and Planning

The improvements to the Plan agent and the introduction of Mermaid diagrams in chat foster better collaboration and understanding. Complex plans or architectural diagrams can be shared and understood instantly, reducing ambiguity and improving team alignment, especially in remote or distributed teams.

How to Upgrade

Upgrading to Visual Studio Code 1.119 is straightforward. As a minor release, it’s recommended for all users, especially those leveraging AI agent capabilities.

Automatic Update: The easiest way to upgrade is to let VS Code update itself. When a new version is available, you’ll typically see a notification. Simply click “Update Now” and restart VS Code.

Manual Download: You can also download the latest version directly from the official VS Code website: https://code.visualstudio.com/download

Choose the installer appropriate for your operating system (Windows, macOS, or Linux). Installing the new version will typically replace your existing installation while preserving your settings and extensions.

For detailed release notes, refer to the official changelog: https://code.visualstudio.com/updates/v1_119