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

🍽️ MenuLocal AI — Regional Menu Translator Powered by Open-Source AI

What I Built

I built MenuLocal AI, a privacy-first, AI-powered regional menu translator for small restaurants, cafés, food stalls, and local food businesses.

Many small restaurants want to serve customers who speak different languages, but translating menus can be expensive. Generic translations can also lose the cultural meaning of regional dishes.

MenuLocal AI helps restaurant owners translate menu items and descriptions into multiple languages while preserving original dish names, prices, and regional food terminology.

Key Features

• 🌍 Translate menus into English, Hindi, Gujarati, Punjabi, and other supported languages.
• 🍛 Preserve original regional dish names alongside their translations.
• ✍️ Generate clear, welcoming, or premium-style menu descriptions.
• 📝 Review and edit translations before publishing.
• 📄 Export translated menus as CSV or Markdown.
• 🔒 Run AI inference locally through Ollama.
• 📴 Support offline inference after installation and model download.

Who would pay for it?

My target customers are independent restaurants, cafés, food stalls, and small restaurant chains that want multilingual menus without relying on a paid cloud AI API.

Potential revenue models include one-time menu translation services, printable menu creation, and a paid desktop version with additional features.

Demo

Status: Local prototype.

🎥 Video demo: Add your demo video link here.

The application can be run locally using Python, Streamlit, and Ollama.

Code

💻 GitHub Repository: [Add your public GitHub repository URL here]

The project uses Python, Streamlit, Ollama, Gemma, Pandas, and Requests.

The repository includes the application code, setup instructions, contribution guidelines, a license, and GitHub issue templates.

Contributions are welcome, especially improvements to regional vocabulary, translation quality, additional languages, and the user interface.

How I Built It

I built MenuLocal AI around an open-weight AI model running locally through Ollama.

Technology stack:

• Python — Application logic.
• Streamlit — Interactive user interface.
• Ollama — Local AI inference.
• Gemma — Default open-weight language model.
• Pandas — Tabular results and CSV exports.
• Requests — Communication with the local Ollama API.
• Pytest — Automated testing.

How it works

1. The restaurant owner enters dish names, descriptions, and prices.
2. They select the source language and target language.
3. The application sends translation instructions to the local AI model.
4. The model generates structured translation results.
5. The user reviews and edits the output.
6. The translated menu can be exported as CSV or Markdown.

The goal is to make regional menu translation accessible to businesses that want more control over their data and AI costs.

Why Does Open Innovation Matter?

Open innovation is central to MenuLocal AI because it gives the project control over its AI model, operating costs, and customization.

🔒 Privacy-first processing

The application can process menu content using a local Ollama instance instead of sending it to a hosted AI API. This provides more control over data handling, provided the application is configured to use a local model.

💰 Cost control

Once the model is downloaded, local inference does not incur a per-request hosted AI API fee. Hardware, electricity, and maintenance still have costs, but avoiding usage-based API charges could make the product more economical for small businesses.

🔄 Model flexibility

Developers can experiment with compatible open-weight models, adjust prompts, and improve regional vocabulary without depending on a single AI provider.

🌐 Offline potential

After installation and model download, the core translation workflow can operate without an internet connection.

🍲 Regional customization

Regional dishes cannot always be translated word for word. Open development allows contributors to improve translations for local cuisines and preserve culturally meaningful dish names.

For this project, open-source AI is more than a technical choice. It creates an opportunity to build a locally operated, customizable product for businesses that need a practical and affordable solution.

Translation quality still needs validation by fluent speakers, especially for dish ingredients, allergens, and dietary claims.

My Agent Session

AI-assisted development helped me plan the architecture, generate the initial implementation, organize the project files, and prepare the documentation.

DevRelay session: Add your saved agent session link here, if available.

Prize Categories

Primary category: Open-Source AI / Side Hustle.

MenuLocal AI explores how an open-weight model and local inference can power a useful product with potential commercial value for small businesses.

I will include additional partner categories only if the project meets their eligibility requirements.

---

Thanks for checking out MenuLocal AI! 🍽️

I'm looking forward to feedback from restaurant owners, developers, and multilingual users.

Contributions that improve translation quality, regional dish-name preservation, accessibility, and usability are especially welcome.