UXForge AI

AI UI/UX feedback team by Sharad Khare.

UXForge AI is a Google ADK multi-agent system that analyzes landing page screenshots, delivers structured UI/UX critique, creates improvement plans, and generates redesigned versions with Gemini image generation.

Built and maintained by Sharad Khare — AI strategist, full-stack developer, and creator of practical agent workflows.

---

What this project does

UXForge coordinates specialized agents to review and improve landing page designs:

1. InfoAgent — answers questions about the system

2. DesignEditor — iteratively refines existing generated designs

3. AnalysisPipeline — sequential pipeline:

- UICritic — comprehensive visual and UX analysis

- DesignStrategist — detailed improvement plan with hex codes and specs

- VisualImplementer — generates improved landing page images

Upload a screenshot and receive critique, strategy, and an improved design.

---

Why Sharad Khare built UXForge AI

Landing page reviews require vision analysis, strategic planning, and visual execution — steps that break when collapsed into one prompt. UXForge demonstrates how a Google ADK coordinator routes to specialists and a sequential pipeline for end-to-end UI/UX feedback with image generation.

---

How it works

User Message + Landing Page Screenshot
   │
   ▼
Coordinator (UIUXFeedbackTeam)
   │
   ├─ General question → InfoAgent
   ├─ Edit existing design → DesignEditor
   └─ New analysis → AnalysisPipeline
         │
         ├─ UICritic → structured feedback
         ├─ DesignStrategist → improvement plan
         └─ VisualImplementer → improved landing page image

Tools

| Tool | Purpose |

|------|---------|

| edit_landing_page_image | Apply targeted edits to an existing design artifact |

| generate_improved_landing_page | Create a new improved design from analysis |

---

Features

• Google ADK coordinator with intelligent routing
• Gemini vision for landing page analysis
• Gemini image generation for improved designs
• Versioned artifact management for design iterations
• Web search for UI/UX best practices and WCAG guidance

---

Quick start

cd uxforge-ai
pip install -r requirements.txt

Set your API key:

export GOOGLE_API_KEY="your_gemini_api_key"

Run the ADK web interface:

adk web

Then select uxforge and upload a landing page screenshot for analysis.

Alternatively, use the runner entrypoint:

adk run app

---

Project structure

uxforge-ai/
├── app.py
├── uxforge/
│   ├── agent.py
│   ├── tools.py
│   └── config.py
├── requirements.txt
├── pyproject.toml
└── README.md

---

Who this is for

• Founders reviewing landing pages before launch
• Designers prototyping AI-assisted critique workflows
• Marketers optimizing conversion-focused pages
• Developers building Google ADK multi-agent products

Use cases

• Get structured UI/UX feedback on a landing page screenshot
• Generate an improved design from critique and strategy
• Iteratively refine generated designs with natural language edits
• Learn coordinator + sequential pipeline patterns with Google ADK

---

Author

• Website: sharadkhare.in

License

MIT © Sharad Khare