This guide covers all features of LumenExMachina in detail.
- Overview
- Creating a Project
- Reference Documents
- Outline Generation
- Outline Chat
- Document Generation
- Illustrations
- Mini-Apps
- Quizzes
- Preview
- Export
- Settings
- Project Files
LumenExMachina follows a step-by-step workflow to create complete teaching materials:
Project Setup → Outline → Documents → Illustrations → Mini-Apps → Quizzes → Export
Each step builds on the previous one. You can always go back and regenerate individual components.
Navigate to the Project tab to set up your teaching material:
| Field | Description | Example |
|---|---|---|
| Topic | The main subject of your material | "Linear Algebra: Vectors and Matrices" |
| Target Audience | Who the material is for | "First-year university CS students" |
| Learning Objectives | What students should learn | "Understand vector operations, matrix multiplication, determinants" |
| Scope | How comprehensive the material should be | Short / Medium / Comprehensive |
- Short — Quick overview, 2-3 chapters, suitable for a single lesson
- Medium — Standard coverage, 4-6 chapters, suitable for a teaching unit
- Comprehensive — In-depth treatment, 7+ chapters, suitable for a full course module
You can upload existing Word (.docx) or PDF files as context for the AI generation. This is especially useful when:
- You have existing teaching materials you want to extend
- You want to ensure alignment with a specific curriculum or textbook
- You have style guidelines the AI should follow
Note: PDF input is supported by Gemini and Anthropic providers. OpenAI and LM Studio do not support PDF input — reference documents will be ignored with these providers.
Upload reference documents in the Project tab before generating the outline.
Click "Generate Outline" to create a hierarchical course structure. The AI generates:
- Chapters (Level 1) with titles and descriptions
- Subchapters (Level 2) with learning objectives
- Sub-subchapters (Level 3) for detailed topics
- Key Terms — Important vocabulary per section
- Illustration Ideas — Suggestions for visual content
- Mini-App Ideas — Interactive learning activity suggestions
The outline serves as the blueprint for all subsequent document generation.
After generating an outline, use the Outline Chat to refine it:
- "Add a chapter about error handling"
- "Make chapter 3 more focused on practical examples"
- "Remove the section about advanced topics"
- "Add more illustration ideas for chapter 2"
The AI updates the outline based on your instructions while maintaining consistency.
Switch to the Documents tab to generate teaching materials. Available document types:
The main learning material. Contains explanations, examples, illustrations, code blocks, and callout boxes. Generated per subchapter with context continuity between chapters.
Quick Mode (Experimental): Enable parallel generation to generate all subchapters simultaneously. Faster, but without cross-chapter context continuity.
Exercises and activities for students. Contains tasks, questions, and guided activities aligned with the outline's learning objectives.
Complete answer key for the worksheet. Generated to match the worksheet's structure.
Self-assessment questions with model answers. Students can check their understanding independently.
Slide-style content suitable for classroom projection. Concise points with key information.
Browser-based quiz with multiple question types (multiple choice, short answer, matching, true/false). Runs as a standalone HTML page.
If a document already exists, clicking "Generate" again will show a confirmation dialog. You can choose to regenerate from scratch or keep the existing version.
The Illustrations tab lets you generate AI images for your documents.
- Click "Generate All" to create images for all illustration ideas from the outline
- Each image is generated in parallel for speed
- Click "Regenerate" on individual images to get a new version
- Images are stored as assets within the project file
Note: Image generation requires the Google Gemini provider. Other providers do not support image generation.
You can select the image model in Settings. The default is models/gemini-3-pro-image-preview. The dropdown shows all available Gemini models, or you can type a custom model name.
Mini-Apps are interactive learning applications embedded in your teaching materials:
- Generate Ideas — AI suggests interactive activities based on your outline
- Create Mini-App — AI generates a complete React or Vanilla JS application
- Preview — Test the mini-app in the built-in browser
Mini-apps are self-contained HTML files that include all necessary JavaScript. They can be:
- Embedded in the document preview
- Exported as standalone HTML files
- Integrated into learning management systems
- React — Uses React 18 + Babel for JSX transpilation. Best for complex interactive UIs.
- Vanilla JS — Pure JavaScript without frameworks. Lighter and simpler.
Generate quizzes in GIFT format for direct import into Moodle LMS:
- Multiple Choice
- True/False
- Short Answer
- Matching
- Numerical
Generate standalone HTML quizzes that run in any browser:
- Immediate feedback per question
- Score tracking
- Explanation display
- No server required
The Preview tab shows a live HTML rendering of your documents:
- Select which document type to preview
- View illustrations inline with text
- Test mini-apps in the embedded browser
- Math formulas rendered with KaTeX
- Code blocks with syntax highlighting
- Fullscreen mode — Open the preview in a separate window
The Export tab provides multiple output formats:
- Professional Word document layout
- Embedded images
- Styled headings, lists, and tables
- Math formulas as images
- Compatible with Microsoft Word, LibreOffice, Google Docs
- Print-ready PDF output
- Generated via QuestPDF
- Embedded images and math formulas
- Consistent typography
- Moodle-compatible quiz format
- Import directly into your LMS
- Supports all standard GIFT question types
- Export individual document types
- Export all documents at once
- Choose output directory
Configure which AI model to use for each task:
- Text Provider — Which AI service to use for text generation (Gemini, Claude, OpenAI, LM Studio)
- Text Model — Specific model for document content
- Outline Provider — AI service for outline generation (can differ from text provider)
- Outline Model — Specific model for outlines
- Image Model — Gemini model for illustration generation
All model dropdowns show available models from the selected provider and are editable for custom model names.
Choose the language for generated teaching materials. This is independent of the UI language — you can use the German interface while generating English materials.
Switch the application interface between English and German.
Projects are saved as .lmx files — ZIP containers containing:
project.lmx
├── project.json # All project data (outline, documents, quizzes, mini-apps, settings)
└── assets/ # Generated images (Base64-encoded PNG files)
Settings (model selection, language preferences) are automatically persisted and restored on app startup.
Projects saved in the older .tfg format can still be loaded but will be saved as .lmx going forward.