Tarim

تارىم

Inspiration

The Islamic Golden Age, which was filled with brilliant minds who established the foundation for modern science, medicine, and mathematics, has always intrigued me. It is quite sad to see this amazing legacy frequently seem trapped in old documents that no one would ever truly touch.

The desire to close that gap became an inspiration for Tarim. In order to make this big impact not only accessible but also engaging and motivating for the younger generation, I wanted to develop a digital platform. I had an idea for a platform that would visually connect ancient discoveries with modern technology, allowing users to feel a direct connection to the researchers who improved the world, alongside other fun features to engage the youth.

What it does

Tarim is an interactive platform that allows users to dive into the Islamic world's scientific and intellectual legacy. It is a collection of related features rather than merely a wiki.

NurAI & Hikmah (ScholarBot): AI-powered chatbots that allow users to get guidance on Islamic topics and "converse" with authentic-sounding representations of historical scholars like Ibn Sina and Al-Khwarizmi.

Scholarpedia & Invention Explorer: Curated and searchable directories of great Muslim scholars and their groundbreaking discoveries and inventions. Users can filter by field, era, and region to explore the vast network of knowledge, or just simply search it up!

Innovation Lineage & Quranic Verse Explorer: Visual tools that promote reflection and exploration by relating Quranic verses (backed by quran.com API) to scientific phenomena and classical inventions to modern day technology.

Personalized Dashboard: a user-focused area where users may save and access their favorite discoveries thanks to features like bookmarking and chat history.

How we built it

The project relies on a strong, modern tech stack. Built with React and Vite, the frontend is a dynamic single-page application with a responsive, clean design thanks to TailwindCSS.

The core of Tarim is a curated PostgreSQL database hosted on Supabase. We spent a good amount of time researching and populating this database with important information on a good amount of key scholars, inventions, and Quranic verses relevant to science. This database acts as our "single source of truth", with the help of AI for tunes and refinement.

We then used AI to enhance this data:

AI for Enrichment: For features like the Scholarpedia modal, the application takes the core, factual data from our database (e.g., a scholar's name, era, and field) and sends it to a backend Supabase Edge Function. This function queries the OpenAI API (GPT-4o), instructing it to generate a rich, narrative-style biography and legacy based on the provided facts. This gives us high-quality, readable content without sacrificing the factual integrity of our own database.

AI-Powered Search and Refinement: The search functionality in the explorers queries our amazing PostgreSQL database to find relevant entries based on the user's keywords and filters. To provide a better experience than just static text, the factual data for each result (e.g., a scholar's name and key works) is then sent to the OpenAI API. We use a carefully engineered prompt to instruct the AI to act as a skilled editor, taking our raw data and transforming it into an engaging, easy-to-read narrative with a consistent tone, which is then presented to the user. This ensures every result is both factually grounded in our database and beautifully made for the best user experience.

Challenges we ran into

Structuring the Data: The first major challenge was designing the database schema. Deciding how to categorize and link scholars, their works, and their inventions in a way that was both accurate and flexible for the frontend was a hard and annoying task that took ages. . Prompt Engineering: It took a lot of iteration to get reliable high-quality results from the AI. Early prompts produced text that was too generic or robotic. To get the desired quality, we had to learn how to write very specific instructions that gave the AI a "persona" (such as "The Curator") and strict formatting guidelines.

Asynchronous Logic: It was hard to integrate the different API calls on the frontend such as quran.com API, AlAdhan API, and others. Using these new API's was quite an experience, whole lot of debugging which was tedious and tiring.

Accomplishments that we're proud of

I'm incredibly proud of building a full-stack application that feels both modern and deeply connected to a rich historical tradition. Specifically, the Hikmah (ScholarBot) feature stands out. Creating very distinct, believable AI personas for historical figures required a deep dive into prompt engineering and feels like a truly unique and new experience.

In addition, the Islamic Calendar page is something I'm especially proud of. Using a Supabase Edge Function, I was able to successfully integrate an Aladhan API, to retrieve precise prayer times based on the user's current geolocation. To create a dynamic and incredibly useful daily dashboard for users, I used the hijri-date library in together with custom logic to calculate and display the exact Hijri date and sun position. I also implemented a custom React hook on the frontend to create a live countdown timer to the next prayer.

What we learned

This project was a fantastic opportunity for learning. The greatest benefit of AI, in my opinion, is not only for content creation but also for refining and bringing to life a foundation of carefully selected factual data. When a well-trained language model and a well-structured database function together smoothly, great things happens.

Additionally, we got a better understanding of React's state management and asynchronous programming. I learned the value of a clear, consistent data flow while developing features that depend on numerous API calls and user interactions.

What's next for Tarim

The vision for Tarim is just beginning, if time permitted, where would have been a lot more features. The next steps are focused on deepening the user's connection to the content:

The Innovation Lineage "Tree" View: Evolving the current timeline into a fully interactive, branching "tech tree" where users can visually trace the lineage of an idea from its ancient roots to multiple modern descendants.

Map Integration: Introducing a map view for the Scholarpedia, allowing users to explore the geographical spread of knowledge across the Islamic world.

Gamification: Makes learning verses competitive or fun. Could be a quiz game: "Guess the Surah" or "Complete the verse". Adds a playful side to a heavy knowledge platform.

Islamic Blog / News Feed: Could be a section called Tarim Insights or Ummah Pulse. Could use AI summarization of Islamic news from curated sources, such as AlJazeera, or other Islamic sources.

Community Features: Potentially adding a feature for users to discuss discoveries and share their own reflections, building a community of curious minds. Make features to make the site more user friendly, a lot of QOL updates.

And much more to come...