Jan 27, 2026

Avalonia DevTools MCP Server

Avalonia DevTools MCP Server

Mike James

Mike James

The new Avalonia DevTools MCP server is available today in 2.2.0-beta2. It's a programmatic bridge between AI agents and your running Avalonia applications, essentially DevTools, but controllable by AI.

The Problem

Large-scale app porting work can be tedious. Whether you're migrating thousands of views, recreating designs from mockups, or fine-tuning layouts to match specifications, the work isn't technically complex; it's just time-consuming and repetitive. The kind of work that burns out good engineers because they're spending their days tweaking margins and alignment instead of solving interesting problems.

The Solution: Programmatic DevTools

The Avalonia DevTools MCP server (`avalonia_devtools`) gives AI agents direct access to inspect and manipulate your running Avalonia applications.

Here's what it provides:

Connection & Discovery

  • Connect to any running Avalonia app

  • List and switch between multiple running instances

Tree Inspection

  • Traverse visual, logical, or merged element trees

  • Search for elements by type or x:Name

  • Walk ancestor chains from any node to root

Property & Style Manipulation

  • Read property values from any UI element

  • Set properties at runtime with full XAML value support

  • Inspect applied styles and setters

  • Query and toggle pseudo-classes (`:pointerover`, `:pressed`, `:focus`, etc.)

Resources & Assets

  • Access resources scoped to the app or specific nodes

  • List embedded assets

  • Download assets via `avares://` URLs

UI Interaction

  • Perform actions: Focus, Enable, Disable, BringIntoView

  • Capture PNG screenshots of any UI element

The Workflow

Here's a typical workflow:

  1. Feed an AI agent a screenshot or design mockup

  2. The agent generates initial Avalonia XAML

  3. Using the MCP server, it connects to the running app

  4. It searches the visual tree to find the elements it created

  5. It inspects property values and compares screenshots

  6. It iterates, adjusting margins, colours, layouts, until it matches

  7. Move to the next view

What would take 30-60 minutes of manual tweaking now takes minutes, with the AI handling the tedious property adjustments while you focus on architecture and logic.

Real-World Example

Last week, I was in Montreal demonstrating this with Devolutions, which is porting its Remote Desktop Manager to Avalonia. They have approximately 5,000 views to migrate, which is a huge amount of repetitive UI work!

We took one of their existing views, gave Claude a screenshot, and let it work. Within a few minutes, a pixel-perfect recreation, running in Avalonia, was ready for testing.

The value isn't just speed; it's the quality of the iteration loop. The agent isn't guessing. It's inspecting, measuring, comparing. It can see issues and fix them proactively. It doesn’t require you to reprompt because it didn’t get it right. It can verify that everything is looking correct before moving on.

This changes the economics of large-scale app porting work!

Available Today

The Avalonia MCP server ships with DevTools 2.2.0-beta2, available now. If you're an Accelerate customer running the latest beta, you can start using it immediately.

The new Avalonia DevTools MCP server is available today in 2.2.0-beta2. It's a programmatic bridge between AI agents and your running Avalonia applications, essentially DevTools, but controllable by AI.

The Problem

Large-scale app porting work can be tedious. Whether you're migrating thousands of views, recreating designs from mockups, or fine-tuning layouts to match specifications, the work isn't technically complex; it's just time-consuming and repetitive. The kind of work that burns out good engineers because they're spending their days tweaking margins and alignment instead of solving interesting problems.

The Solution: Programmatic DevTools

The Avalonia DevTools MCP server (`avalonia_devtools`) gives AI agents direct access to inspect and manipulate your running Avalonia applications.

Here's what it provides:

Connection & Discovery

  • Connect to any running Avalonia app

  • List and switch between multiple running instances

Tree Inspection

  • Traverse visual, logical, or merged element trees

  • Search for elements by type or x:Name

  • Walk ancestor chains from any node to root

Property & Style Manipulation

  • Read property values from any UI element

  • Set properties at runtime with full XAML value support

  • Inspect applied styles and setters

  • Query and toggle pseudo-classes (`:pointerover`, `:pressed`, `:focus`, etc.)

Resources & Assets

  • Access resources scoped to the app or specific nodes

  • List embedded assets

  • Download assets via `avares://` URLs

UI Interaction

  • Perform actions: Focus, Enable, Disable, BringIntoView

  • Capture PNG screenshots of any UI element

The Workflow

Here's a typical workflow:

  1. Feed an AI agent a screenshot or design mockup

  2. The agent generates initial Avalonia XAML

  3. Using the MCP server, it connects to the running app

  4. It searches the visual tree to find the elements it created

  5. It inspects property values and compares screenshots

  6. It iterates, adjusting margins, colours, layouts, until it matches

  7. Move to the next view

What would take 30-60 minutes of manual tweaking now takes minutes, with the AI handling the tedious property adjustments while you focus on architecture and logic.

Real-World Example

Last week, I was in Montreal demonstrating this with Devolutions, which is porting its Remote Desktop Manager to Avalonia. They have approximately 5,000 views to migrate, which is a huge amount of repetitive UI work!

We took one of their existing views, gave Claude a screenshot, and let it work. Within a few minutes, a pixel-perfect recreation, running in Avalonia, was ready for testing.

The value isn't just speed; it's the quality of the iteration loop. The agent isn't guessing. It's inspecting, measuring, comparing. It can see issues and fix them proactively. It doesn’t require you to reprompt because it didn’t get it right. It can verify that everything is looking correct before moving on.

This changes the economics of large-scale app porting work!

Available Today

The Avalonia MCP server ships with DevTools 2.2.0-beta2, available now. If you're an Accelerate customer running the latest beta, you can start using it immediately.

The new Avalonia DevTools MCP server is available today in 2.2.0-beta2. It's a programmatic bridge between AI agents and your running Avalonia applications, essentially DevTools, but controllable by AI.

The Problem

Large-scale app porting work can be tedious. Whether you're migrating thousands of views, recreating designs from mockups, or fine-tuning layouts to match specifications, the work isn't technically complex; it's just time-consuming and repetitive. The kind of work that burns out good engineers because they're spending their days tweaking margins and alignment instead of solving interesting problems.

The Solution: Programmatic DevTools

The Avalonia DevTools MCP server (`avalonia_devtools`) gives AI agents direct access to inspect and manipulate your running Avalonia applications.

Here's what it provides:

Connection & Discovery

  • Connect to any running Avalonia app

  • List and switch between multiple running instances

Tree Inspection

  • Traverse visual, logical, or merged element trees

  • Search for elements by type or x:Name

  • Walk ancestor chains from any node to root

Property & Style Manipulation

  • Read property values from any UI element

  • Set properties at runtime with full XAML value support

  • Inspect applied styles and setters

  • Query and toggle pseudo-classes (`:pointerover`, `:pressed`, `:focus`, etc.)

Resources & Assets

  • Access resources scoped to the app or specific nodes

  • List embedded assets

  • Download assets via `avares://` URLs

UI Interaction

  • Perform actions: Focus, Enable, Disable, BringIntoView

  • Capture PNG screenshots of any UI element

The Workflow

Here's a typical workflow:

  1. Feed an AI agent a screenshot or design mockup

  2. The agent generates initial Avalonia XAML

  3. Using the MCP server, it connects to the running app

  4. It searches the visual tree to find the elements it created

  5. It inspects property values and compares screenshots

  6. It iterates, adjusting margins, colours, layouts, until it matches

  7. Move to the next view

What would take 30-60 minutes of manual tweaking now takes minutes, with the AI handling the tedious property adjustments while you focus on architecture and logic.

Real-World Example

Last week, I was in Montreal demonstrating this with Devolutions, which is porting its Remote Desktop Manager to Avalonia. They have approximately 5,000 views to migrate, which is a huge amount of repetitive UI work!

We took one of their existing views, gave Claude a screenshot, and let it work. Within a few minutes, a pixel-perfect recreation, running in Avalonia, was ready for testing.

The value isn't just speed; it's the quality of the iteration loop. The agent isn't guessing. It's inspecting, measuring, comparing. It can see issues and fix them proactively. It doesn’t require you to reprompt because it didn’t get it right. It can verify that everything is looking correct before moving on.

This changes the economics of large-scale app porting work!

Available Today

The Avalonia MCP server ships with DevTools 2.2.0-beta2, available now. If you're an Accelerate customer running the latest beta, you can start using it immediately.

Continue Reading

The latest News articles