BrandSpark: AI Marketing Suite for Small Businesses and Entrepreneurs

This is a submission for the Google AI Studio Multimodal Challenge

What I Built

I built BrandSpark, a comprehensive, AI-powered marketing suite designed to empower small businesses, e-commerce sellers, and solo entrepreneurs. It solves the common challenge of creating high-quality marketing assets without a large budget or a dedicated creative team.

BrandSpark transforms a simple product photograph into a complete set of marketing materials through three core features:

1. AI Photo Shoot: Users can upload a product image and instantly generate professional-grade photos with custom backgrounds, lighting, and styles, turning a basic shot into a studio-quality image.
2. AI Ad Creator: It creates stunning, ready-to-post social media ads by artistically integrating promotional text directly onto the product image, matching a selected design style.
3. AI Campaign Planner: It generates a complete 7-day social media marketing strategy, including daily themes, captions, hashtags, and calls to action, all tailored to the specific product and a chosen campaign goal.

Demo

How I Used Google AI Studio

I leveraged the multimodal capabilities of the Gemini API to power all of BrandSpark's core features. The application makes extensive use of providing both image and text data in prompts to generate nuanced, context-aware image and text outputs.

1. gemini-2.5-flash-image-preview is the engine behind all visual generation. I used it for the AI Photo Shoot and the Social Post Creator. Its ability to interpret a text prompt and an input image to perform complex edits—like changing the background, adjusting lighting, and graphically adding text—is central to the app's functionality.
2. gemini-2.5-flash is used for all text and data generation tasks. Its multimodal understanding is key, as it analyzes the input image to inform its text output.
   • Scene Suggestion: Takes an image and a text instruction to generate a creative text prompt.
     • AI Copywriter & Campaign Planner: These features use multimodal input (image + text prompt) combined with JSON Mode (responseSchema). This was crucial for getting reliable, structured data back from the API. By defining a schema, the application can predictably parse the generated headlines, captions, and full campaign plans into a structured UI without complex string manipulation.

Multimodal Features

BrandSpark is built from the ground up on multimodality, combining image and text understanding to create a tool that is more than the sum of its parts.

1. Context-Aware Image Generation (Image + Text -> Image): Instead of just generating images from text, the app edits existing images based on text. This enhances user experience by giving them creative control to perfect their actual product photos. A user can guide the AI to place their product in a "rustic wooden surface with warm, natural feel" or a "minimalist concrete slab," achieving professional results instantly without a physical photoshoot.
2. Visually-Grounded Content Strategy (Image + Text -> Structured Text/JSON): This is the most powerful multimodal feature. The AI doesn't just write generic marketing copy; it looks at the product in the image to generate its strategy. When it plans a campaign for a coffee maker, the generated captions, themes, and hashtags are all relevant to coffee, morning routines, and kitchen gadgets. This visual grounding makes the generated marketing content far more specific, relevant, and effective, saving the user hours of brainstorming.

By combining these modalities, BrandSpark transforms a single, simple product photo into a complete, ready-to-launch marketing campaign.