LocalBiz AI

Turn one product into a complete AI-powered marketing campaign in seconds.

Comment

Inspiration

Small businesses often rely on social media to reach customers, but creating marketing content consistently takes time. Business owners already handle products, orders, deliveries, and customer service. Writing captions, designing promotional images, and deciding when to post adds another responsibility.

I wanted to make that process easier, especially for local businesses without a dedicated marketing team.

That inspired LocalBiz AI: a marketing workspace that turns saved product information and photos into editable social media drafts, promotional images, and a posting plan.

What it does

LocalBiz AI brings the main steps of creating a social media campaign into one workspace.

Business owners can:

  • Create an account and save their business profile.
  • Add, edit, and delete products with descriptions, prices, and product photos.
  • Generate campaigns for Facebook or Instagram using a selected product, campaign goal, target audience, tone, and language.
  • Choose English, Sinhala, or Tamil for campaign drafts.
  • Receive three post ideas containing captions, calls to action, hashtags, and visual suggestions.
  • Review and edit the generated content, copy post text, and save campaigns for later.
  • Generate AI promotional posters through deAPI using campaign content, an editable creative prompt, and an optional product reference photo.
  • Reuse a saved product photo or upload a different reference for an individual poster.
  • Create free branded photo posters as an alternative to AI image generation.
  • Save and download promotional images as 1080 × 1080 PNGs.
  • Organize posting dates in a calendar, reschedule or cancel plans, and manually mark shared posts as published.

Campaign drafts use the saved business context, product facts, and selected tone. Owners remain in control by reviewing the text and images before sharing.

Posting dates guide manual sharing. The current version does not automatically publish to social media.

How we built it

I built LocalBiz AI as a full-stack web application during LovHack Season 3.

The frontend uses React, Vite, Tailwind CSS, and React Router to provide a responsive workspace for desktop and mobile users. The backend uses Node.js and Express, with MongoDB Atlas storing account data, business profiles, products, campaigns, photos, posters, and posting plans.

Google Gemini generates structured campaign drafts from the business profile, selected product, and campaign brief. The backend validates the generated output before returning it to the user.

deAPI powers AI promotional image generation. The integration uses Qwen Image Edit Plus for requests with reference photos and FLUX.2 Klein for requests without a reference. Image jobs run asynchronously, with progress checking and recovery so users can return to an existing job without submitting another paid generation.

The free poster builder uses browser canvas, while Sharp handles image validation and processing on the backend.

The frontend is deployed on Vercel, and the backend is hosted on Render. API keys remain on the backend, and saved data is scoped to the authenticated account.

The main workflow is:

Business Profile → Product and Photo → Campaign Brief → AI Drafts → Review and Edit → Promotional Poster → Posting Plan

Challenges we ran into

One challenge was generating useful campaign drafts while keeping product details, prices, and business information accurate. I used structured prompts, output validation, and editable drafts to support that workflow.

AI image generation introduced another challenge: a reference photo can guide the result, but the model may still change product details or produce imperfect text. I strengthened the reference instructions and kept image review central to the experience.

Managing paid image requests also required careful handling. I added price checks, generation limits, and resumable jobs to reduce accidental spending and duplicate submissions.

Deployment involved configuring environment variables, database access, frontend-to-backend routing, and production sessions across Vercel and Render.

Throughout the build, I kept the scope focused on a working marketing workflow that could be demonstrated within the hackathon period.

Accomplishments that we're proud of

  • Built and deployed a complete workflow connecting products, campaign drafts, promotional images, and a content calendar.
  • Integrated Gemini for campaign text and deAPI for AI promotional posters.
  • Added private product photo uploads that can be reused as image-generation references.
  • Created both AI image generation and a free photo-template option.
  • Implemented account-based access to saved business data, campaigns, and images.
  • Verified core behavior with 90 automated tests using a disposable database and simulated AI providers, alongside desktop and mobile browser checks.

What we learned

This project taught me how to integrate generative AI into a practical application with saved data, editable results, and a clear user workflow.

I learned that useful AI features need more than a prompt. They also need validation, cost controls, understandable errors, and ways to recover from slow or failed requests.

I also learned that AI-generated text and images require different approaches. Business context helps guide captions, while image generation needs clear visual instructions and careful review.

Building and deploying the application gave me experience with authentication, private data access, image processing, databases, and production configuration.

What's next for LocalBiz AI

Future improvements include:

  • Better product fidelity and cleaner typography in AI posters.
  • Brand kits with saved logos, colors, fonts, and reusable layouts.
  • Additional content formats for WhatsApp, TikTok, and Reels.
  • Direct social media publishing with connected accounts.
  • Campaign performance analytics.
  • Stronger language-quality checks and brand-voice customization.
  • More scalable image storage and account management.

The long-term goal is to make useful digital marketing tools accessible to small businesses with limited time, budget, and resources.

One product. Editable content. Promotional images. A clear posting plan.

Built With

Share this project:

Updates

Submission history