Overview
Chat with Classic is a web-based AI assistant designed to make ancient Chinese literature more accessible. Traditional texts are often difficult to read, featuring complex character structures, vertical layouts, and outdated formats. This project leverages OCR (Optical Character Recognition) and AI-powered summarization to simplify reading, enhance research efficiency, and provide interactive engagement with historical texts.
With Chat with Classic, users can:
📖 Convert scanned ancient texts into modern digital formats.
🤖 Interact with an AI-powered chatbot for deeper understanding.
📝 Organize literature into a personalized knowledge base for research and writing.
This project was developed as part of the 2023 UXD AWARD, aligning with the theme of using AI for a better human experience.
My contribution
Product strategy
UI/UX design
The team
1 x developer
1 x designer
Year
Jun 2023 - Aug 2023
Process
Understanding the Challenges
Through interviews with students, professors, librarians, and curators, we identified key pain points in handling ancient texts:
1️⃣ Reading Complexity – Many historical texts use traditional or rare characters, making them hard to decipher.
2️⃣ Inefficient Excerpting – Manually summarizing key information is time-consuming.
3️⃣ Fragmented Research – Existing tools lack a structured system to manage research materials.
Our goal was to develop a seamless, AI-assisted workflow to address these issues.
Designing the Solution
The platform was designed around four core functionalities:
✔ OCR-powered Text Conversion – Using Xunfei OCR API, the system accurately digitizes ancient texts.
✔ AI Chat for Literature Interpretation – ChatGPT-4, integrated with LangChain, allows users to discuss and summarize texts.
✔ Knowledge Base Management – Users can create flashcards and reference notes, making research more efficient.
✔ Translation & Formatting Tools – The system converts traditional Chinese to modern formats and supports basic English translations.
Collaboration & Development
I collaborated with a developer to implement the backend using Python, Streamlit, and API integrations.
Key technical challenges and solutions:
💡 OCR Accuracy – We refined the recognition pipeline to reduce misinterpretations.
🚀 AI Response Speed – Limited text input per request to improve chatbot efficiency.
🎨 User Experience – Designed a clean, research-friendly interface, tested with eye-tracking studies to refine usability.
Outcome
The prototype successfully streamlined literature research, receiving high user satisfaction ratings:
⭐ 85% of testers found the AI chat helpful for summarizing complex texts.
⭐ Users reported a 50% reduction in research time (saving 1-2 hours per day compared to manual methods).
⭐ The project scored 32/35 in usability tests, with praise for its clean interface and innovative AI use.
User feedback led to key improvements:
🔹 Added side-by-side text comparison for better proofreading.
🔹 Optimized translation speed to enhance workflow efficiency.
🔹 Introduced customizable font sizes for better readability.
Chat with Classic bridges the gap between historical literature and modern AI, making ancient texts more accessible, searchable, and interactive for researchers and enthusiasts.
💁🏻♀️ View detailed case study —>