Morph: AI Browser Extension Builder

Inspiration

Browser extensions are powerful, but building one still requires knowing JavaScript, Chrome APIs, manifests, permissions, and web-page structure. We wanted to make that power accessible to anyone with a useful browser problem to solve.

Morph was built around a simple question: what if you could describe a browser improvement in plain language and get a working extension back?

What it does

Morph is an AI browser extension builder. Users describe what they want—such as making articles easier to read, hiding distracting content, improving accessibility, or adding helpful controls—and Morph generates a Chrome extension for it.

It can:

Turn a natural-language request into extension code
Create Manifest V3 files and content scripts
Validate the generated extension
Build and package it for Chrome
Help users install and test it
Iterate when the first version needs improvement

The goal is not just to generate code, but to help people solve real browsing problems with something they can actually use.

How we built it

Morph uses a React and Vite interface for the extension experience, a FastAPI backend for orchestration, and WebSockets for live progress updates while an extension is being created.

The AI coding workflow can inspect project files, create and edit extension code, run validation, build the extension, and load it for testing. Generated extensions use Chrome Manifest V3 and standard browser APIs, keeping the output portable and understandable.

Challenges we ran into

The hardest part was making generated extensions reliable on real websites. Every site has different HTML structures, naming conventions, dynamic content, and navigation behavior.

We also had to handle:

Pages that change after loading
Content rendered by JavaScript
Selectors that break when sites update
Chrome permissions and extension lifecycle rules
WebSocket connections closing during background updates
Clear errors when a project cannot be loaded or built

A generated extension is only useful if it works outside a controlled demo, so we tested realistic scenarios on public websites and repaired failures instead of treating generation alone as success.

Accomplishments that we're proud of

We built a working end-to-end flow from plain-language request to installable Chrome extension.

We are especially proud that Morph:

Generates complete Manifest V3 extensions
Provides live build and validation feedback
Supports iterative fixes
Produces extensions for practical use cases
Includes a polished, focused interface
Was evaluated across 20 realistic browser-extension scenarios
Handles common backend and browser-runtime failure cases clearly

Morph makes browser customization feel closer to explaining a task to a teammate than learning an entirely new development stack.

What we learned

We learned that code generation is only the beginning. The real product is the feedback loop: generate, validate, run, observe, repair, and test again.

We also learned that browser automation requires humility. A solution that works on one page may fail on another, and a successful build does not guarantee useful behavior. Reliable generation requires strong validation, visible errors, and testing against real websites.

What's next for Morph: Your AI Browser Extension Builder

Next, we want to make Morph more reliable, more accessible, and easier to share.

Planned improvements include:

Better support for dynamic and single-page websites
Built-in accessibility-focused extension templates
Safer permission recommendations
More transparent explanations of generated code
Support for additional AI providers and local models
One-click extension updates and version history
Stronger privacy controls for page content
A community library for discovering and remixing extensions
Automatic regression testing when websites change

Our long-term goal is to make Morph a practical layer between people and the web—helping anyone shape their browsing experience without needing to become an extension developer first.