EduAI

EduAI is a VS Code extension for project-based frontend learning.

It analyzes project code, retrieves relevant code chunks from natural language questions, and helps generate:

• code explanations
• practice exercises
• practice grading feedback
• AI-generated learning material (Markdown + HTML + PDF)

Features

• Project analysis: detects framework, structure, and code patterns.
• Natural-language code Q&A: understands questions and finds relevant project code snippets.
• Practice mode: creates lightweight exercises and evaluates submitted answers.
• Learning material export: generates study material files in eduai-output/.

Requirements

• VS Code ^1.110.0
• Local Ollama server running at http://localhost:11434
• Ollama model (default: qwen2.5-coder:7b)

OpenAI key setup

For Marketplace users, register your own OpenAI API key from the Command Palette:

• EduAI: OpenAI API 키 등록
• EduAI: OpenAI API 키 삭제

Extension settings

You can customize EduAI in VS Code Settings:

• eduai.ollamaUrl (default: http://localhost:11434/api/chat)
• eduai.ollamaModel (default: qwen2.5-coder:7b)
• eduai.ollamaFallbackModel (default: ``; optional timeout fallback model)
• eduai.retrievalTopK (default: 5)
• eduai.maxProjectFiles (default: 150)
• eduai.referenceKnowledgeEnabled (default: true)
• eduai.referenceSourceUrls (default: React.dev + MDN 기본 URL)
• eduai.referenceMaxPages (default: 4)
• eduai.referenceCacheTtlHours (default: 168)

When reference knowledge is enabled, EduAI uses trusted public sources (react.dev, developer.mozilla.org) and includes a "참고 출처" section in generated learning materials.

How to use

1. Open the EduAI Chat view in the activity bar.
2. Click 프로젝트 분석.
3. Ask a question in natural language.
4. (Optional) Generate practice and submit an answer for feedback.
5. Click 학습자료 생성 to export Markdown/HTML/PDF material.

Output files

Generated files are saved under:

• eduai-output/eduai-material-*.md
• eduai-output/eduai-material-*.html
• eduai-output/eduai-material-*.pdf

Known limitations

• Output quality depends on local model capability and prompt adherence.
• PDF 렌더링은 로컬 Chrome/Edge/Chromium 실행 가능 환경에 의존합니다.