Elton Kadiu

SketchShip AI is a web application that converts hand-drawn napkin sketches and UI wireframes into fully executable, running Streamlit applications in seconds — with zero manual coding required from the user. The user uploads a photo of their sketch (or draws one on paper and takes a phone photo). The app sends the image to Meta's Muse Spark model, which reads the sketch and returns a structured JSON layout describing every UI component it detects: titles, text inputs, number fields, dropdowns, buttons, tables, charts, and the overall layout structure. That JSON is then sent back to Muse Spark, which writes clean, executable Python Streamlit code from the visual description. The generated code is statically validated (import allow-list, blocked dangerous calls like eval and open) and executed live inside the app, giving the user an interactive preview they can click and type in immediately. The user also gets the raw generated code with syntax highlighting and a one-click download button, so they can take the app.py file and deploy it anywhere. The entire pipeline — image understanding, code generation, live preview — runs in under 60 seconds with no local GPU, no paid infrastructure beyond the Meta Model API credits, and no coding knowledge required from the user. This solves a real problem: the gap between having an idea and having a working prototype is too large for most people.SketchShip AI closes that gap completely. Draw it, ship it. Tech stack: Python 3.11, Streamlit, Meta Model API (Muse Spark 1.3), Pillow, python-dotenv, requests.