Upload Your Project
Drop a ZIP archive or individual files. Full folder structure is preserved and analyzed recursively.
Drop ZIP or files here
Supports entire project directories via ZIP.
Or drag individual HTML, CSS, JS, image, and font files.
Or drag individual HTML, CSS, JS, image, and font files.
.zip
.html
.css
.js / .ts
.jsx
.svg
.png .jpg .webp
.woff .ttf
How it works
1
Upload your project
ZIP your entire project folder or drop files individually. All assets, styles, and scripts are detected.
2
Deep analysis & architecture map
The system parses every file, builds a dependency tree, detects CDN references, and generates an interactive visual map.
3
Configure AI models
Add API keys for Mistral, Grok, Hugging Face, or others. The orchestrator auto-routes tasks by complexity.
4
AI converts to Shopify Liquid
Every color, string, image, and layout value is extracted into a {% schema %} block with correct input types.
5
Copy & paste to Shopify
The output is production-ready. Follow the manifest for asset uploads and theme editor testing.
Architecture Map
Dependency tree + file interconnections + asset usage
Sandboxed live preview
Upload &
Parse
Parse
Deep
Analysis
Analysis
Arch
Map
Map
Schema
Extract
Extract
AI
Refactor
Refactor
Validate
Liquid
Liquid
Final
Output
Output
AI Orchestration
Mistral 7B โ Light tasks
idle
Parsing ยท Syntax analysis ยท String extraction
Tokens used: 0
Rate: โ/min
Mistral Large โ Heavy tasks
idle
Schema mapping ยท Logic refactor ยท Liquid generation
Tokens used: 0
Context: โ
HuggingFace โ Fallback
standby
Rate-limit failover ยท Parallel processing
Upload files and configure API keys to begin
Task Queue & Log
โWaiting for pipeline to start...
Complexity Scores
Run analysis to compute complexity scores
Liquid Output
Ready-to-paste Shopify section file
section.liquid
โ
Run the pipeline to generate Liquid code...
API Settings
Keys are stored encrypted in memory only. Never sent to any server except the respective AI provider.
๐ All keys are stored in-session only and never persisted to disk or transmitted to third parties.
๐
Mistral AI
not configured
Used for: light parsing, syntax analysis, string extraction (mistral-7b-instruct) and heavy logic refactoring (mistral-large-latest)
๐
Grok (xAI)
not configured
Used for: complex schema reasoning, context-heavy transformations
๐ค
Hugging Face Inference
not configured
Used for: fallback when primary models hit rate limits; parallel processing
๐
Custom OpenAI-compatible Endpoint
not configured
Any OpenAI-compatible API (Cerberus, local LLMs, Together AI, etc.)
โ
Orchestration Settings
5