MarketMate AI

This is a submission for the [Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle]

What I Built

I built MarketMate AI, a marketing assistant designed for small businesses, freelancers, and creators who need practical marketing help but may not have the budget to hire a marketing agency.

Many small businesses struggle with questions like: Who are my potential customers? Which marketing channels should I try? What should I post this week? How do I reach potential customers professionally?

MarketMate AI turns basic business information into a starting point for answering those questions.

It includes four tools:

• Audience Finder: Suggests potential customer segments and ways to reach them.
• Marketing Strategy Builder: Creates a positioning statement, possible differentiators, marketing channels, and a seven-day action plan.
• Content and Campaign Studio: Generates social media ideas, captions, calls to action, and a content calendar.
• Customer Outreach Writer: Drafts introductions, follow-ups, and value-first outreach messages.

The intended users are small business owners, independent creators, and freelancers who want an affordable starting point for their marketing work.

Demo

https://jee-grid.site

Code

GitHub repository: https://github.com/knull-001/MarketMate-AI

The project uses Python and FastAPI for the backend, with HTML, CSS, and JavaScript for the frontend.

How I Built It

I built MarketMate AI using:

• Python and FastAPI for API endpoints
• HTML, CSS, and JavaScript for the user interface
• Google's Gemini API for the current text-generation provider
• GitHub Codespaces for browser-based development

I connected Gemini using Google's Python SDK and kept the API key in an environment variable rather than committing it to GitHub.

One challenge was debugging why the app kept returning demo responses after the API integration. I checked the provider configuration and corrected the backend's provider reporting so I could distinguish Gemini responses from demo-mode responses.

An important limitation: the current working version uses the hosted Gemini API. Gemini is not an open-weight model, so I won't claim that the current generation pipeline is open-source AI. Integrating an open-weight model such as Gemma as a working provider is the next step I need to complete for this challenge.

Why Does Open Innovation Matter?

For a small-business tool, flexibility matters. Different users have different budgets, privacy requirements, hardware, and internet access.

An open-weight model could give this project more control over model choice, deployment, and customization. It could also make it possible to run inference in an environment chosen by the user, depending on the model's hardware requirements and license.

The current Gemini integration helped me get a working prototype running, but it doesn't provide those open-model benefits by itself. I want to make the open component a real part of how MarketMate AI works, rather than treating an open-source code repository as proof that its AI model is open source.