Acadexis: AI-Driven Personalized Learning Platform for Universities

Acadexis is an institutional-level AI knowledge base platform for universities, connecting teachers, students, and AI learning tools. Its core design principle is 'knowledge anchoring'—AI answers must be based on verified course materials uploaded by lecturers, with precise page citations to avoid hallucinations. This platform combines AI's interactive advantages with academic rigor, solving the problem of generic AI tools in academic scenarios. It is an open-source project maintained by oluwaseyipd on GitHub (project link: https://github.com/oluwaseyipd/Acadexis_frontend, released on June 2, 2026).

With the rapid development of large language models, education is undergoing profound changes. However, generic AI chat tools have obvious limitations in academic scenarios: they may produce 'hallucinations' (unverified information) and cannot deeply integrate with specific course content. Acadexis was developed to address these issues, building an AI-enhanced education platform based on course materials as the knowledge foundation.

Core Design

Acadexis follows the 'knowledge anchoring' principle: all AI answers must reference uploaded PDF materials with precise page numbers, and lecturers fully control the AI's knowledge boundary.

Student Features

• Personalized dashboard: Visualized learning statistics, recent activity tracking, course progress overview.
• Study Lab: AI dialogue interface, intelligent Q&A based on uploaded materials, precise page citations.
• Quiz system: Timed assessments, detailed score analysis, wrong question review.
• Library: Access to course materials, PDF/document downloads.

Lecturer Features

• Teaching dashboard: Overview of taught courses, student analysis data.
• Knowledge Hub: Upload course materials, organize documents, build AI knowledge bases.
• Learning difficulty heatmap: Real-time analysis of AI interaction data to identify confusing knowledge points.
• Student management: Track registered students, manage permissions.
• Quiz creation: Design course quizzes, manage question banks.

Tech Stack

• Framework: Next.js 16.2.4 (App Router)
• Language: TypeScript 5.x
• UI: React 19.2.4, Tailwind CSS 3.4.x, Radix UI (accessible components)
• State management: Zustand 5.x
• Form handling: React Hook Form + Zod
• Others: Axios (HTTP client), Recharts (charts), Framer Motion (animations), Lucide React (icons)

Key Technical Innovations

1. AI Knowledge Anchoring: Different from traditional RAG, it requires AI answers to cite uploaded PDF materials with page numbers, controlled by lecturers.
2. Real-time Analysis: The 'learning difficulty heatmap' identifies frequently asked concepts, materials needing supplements, and weak points in course design.
3. Modern Frontend: Next.js App Router uses server components (reduces client JS size), streaming (faster first load), nested layouts, and server actions (simplifies data changes).
4. Accessibility: Radix UI ensures keyboard navigation, screen reader compatibility, focus management, and correct ARIA attributes.

Scenario 1: Large Open Course

For a 500-student 'Introduction to AI' course:

• Lecturers upload courseware, papers, and references to the Knowledge Hub.
• Students use the Study Lab to ask questions, with AI answers based on materials.
• The system identifies confusing chapters via heatmap.
• Lecturers adjust teaching focus based on the heatmap.

Scenario 2: Graduate Mentor Guidance

• Mentors create exclusive course spaces for each student.
• Upload relevant papers and references.
• Students discuss with AI based on specified literature.
• Mentors view students' question history to understand their thinking process.

Scenario3: Enterprise Training

• Upload company policies, product manuals.
• New employees use AI Q&A to quickly familiarize with the company.
• Ensures all answers are based on official documents, avoiding information bias.

Future Direction of EdTech

Acadexis represents an important direction for educational AI: evolving from generic dialogue to domain-specific, controllable, and verifiable intelligent tutoring. Key takeaways:

1. Importance of knowledge boundaries: AI capabilities should be limited to verifiable knowledge.
2. Necessity of citation mechanisms: Academic scenarios require traceable information sources.
3. Data-driven teaching: Learning analysis should feed back to teaching improvement in real time.
4. Role-differentiated design: Students and teachers need different functional perspectives.

Conclusion

Acadexis is a well-designed, technologically advanced AI education platform. It successfully combines LLM interaction capabilities with academic rigor, solving the AI hallucination problem via knowledge anchoring and providing teaching insights through real-time analysis. It is an open-source project worth studying for institutions and developers exploring AI-enabled education.