Mrj AI Studio Pro

From Photo to Campaign, Studio-Level Creativity in Minutes.

Comment

Inspiration

As a senior creative designer, I saw how small businesses struggle to produce high-quality marketing visuals without expensive teams. I wanted to build a tool that gives studio-level creative power to anyone using Gemini 3 multimodal intelligence. I’m not a traditional developer, so this project represents deep personal effort and many late nights learning, testing, and rebuilding.

What it does

MrJ AI Studio Pro turns raw product photos into campaign-ready marketing visuals. It analyzes images, understands context, and generates high-quality scenes plus aligned headlines and CTAs.

How we built it

Built with Gemini 3 API using multimodal inputs (image + reference + intent). Gemini Flash handles Smart Visual Scan and structured JSON reasoning. Gemini Pro Image handles high-fidelity synthesis and generative editing. The pipeline is:

𝐶𝑟𝑒𝑎𝑡𝑖𝑣𝑒 = 𝐺𝑒𝑚𝑖𝑛𝑖𝑃𝑟𝑜(𝑅𝑒𝑓𝑖𝑛𝑒𝑑𝑃𝑟𝑜𝑚𝑝𝑡𝐹𝑙𝑎𝑠ℎ(𝐼𝑚𝑎𝑔𝑒+𝐼𝑛𝑡𝑒𝑛𝑡)) React + Google GenAI SDK + client-side storage power the app.

Challenges we ran into

Learning API behavior, quota limits, retries, and model routing without a formal dev background was the hardest part. Many failures, overload errors, and rebuild cycles.

Accomplishments that we're proud of

Working multimodal routing, schema outputs, and generative isolation without third-party tools.

What we learned

Gemini reasoning > prompt guessing. Structure beats randomness.

What's next for Mrj AI Studio Pro

Smarter automation, batch campaigns, and deeper creative reasoning workflows.

Built With

Share this project:

Updates