-
-
Dashboard showing today's lessons, upcoming appointments, and conflicts
-
Interactive week view with click-to-create functionality for lessons and blocked times
-
Conflict detection and blocked time management in the week view
-
Invoice list and detail view with automatic generation from taught lessons
Inspiration
As a private tutor, I often experienced the same recurring problem: scattered information, inconsistent scheduling, missing overview of income, and no tool that can handle real-world tutoring workflows — including contracts, travel times, recurring lessons, conflicts, billing, and AI-assisted lesson planning.
Existing tools are either too simple (pure calendars) or too complex (full CRM systems). I wanted something practical, fast, ethical, and secure — especially because tutoring involves minors and personal data. This became the motivation for TutorFlow: a structured, privacy-conscious management tool for tutors.
What it does
TutorFlow is a Django-based web application that allows tutors to: • Manage students, contracts, planned hours, and monthly quotas • Schedule lessons in a weekly calendar with conflict detection • Automatically track income (taught / invoiced / paid) • Generate invoices and mark included lessons as paid • Block time for personal events (vacation, university, work) • Use an LLM-powered lesson plan generator (with strict privacy controls) • Run a deterministic demo mode (MOCK_LLM) for safe hackathon evaluation
TutorFlow supports English/German, includes a one-command demo setup, and offers a clear and secure workflow designed for real tutors.
How we built it
TutorFlow is built with modern Django 6.0 and a modular architecture: • Backend: Django + Django ORM + modular apps (students, lessons, billing, ai, etc.) • Frontend: Django templates, HTMX-style dynamic updates, custom JS/CSS (no inline scripts, CSP-compliant) • AI Layer: pluggable LLMClient with optional mock-mode (MOCK_LLM=1) • Security: CSP headers, PII sanitizer, strict ENV-based secrets, mock defaults • Recurring Scheduling Engine: weekly / biweekly / monthly rules • Conflict Engine: time conflicts + monthly quota checks + blocked-time overlaps • Billing Engine: transactional invoice creation, reversal logic, correct unit-pricing • Testing: unit tests + integration tests + CI with GitHub Actions
The project is fully internationalized, documented, and structured for maintainability. All demo content is deterministic and free of personal data.
Challenges we ran into
• Calendar UI complexity: implementing a weekly teaching calendar with drag-to-create, conflict overlays, and correct time placement required several refactorings. • Accurate conflict detection: quota-based conflicts, time overlaps, and recurring events interacting cleanly was surprisingly intricate. • Secure AI integration: preventing personal data from entering prompts required a custom sanitization layer and a mock-LLM system. • Invoice correctness: correctly handling 45-minute units, taught vs. planned vs. invoiced states, and rollback on invoice deletion needed careful transactional design. • Strict CSP: removing all inline JS/CSS across templates took effort but greatly improves security. • Demo reproducibility: guaranteeing that “one command” fully sets up the system, including fake LLM responses and demo accounts.
Accomplishments that we're proud of
• Fully working tutor management system built from scratch during the hackathon • Secure AI integration with sanitization + mock-mode (no external calls required) • Clean modular architecture with documented diagrams • Internationalization (EN/DE) and deterministic demo • Polished UI with a week-calendar and conflict visualization • Robust invoice workflow with automatic state changes • CSP-compliant frontend with no inline scripts • Comprehensive documentation
What we learned
• Good scheduling logic is harder than it looks — real-world time planning requires edge cases. • AI integration must be privacy-first, especially when dealing with minors. • Mock systems are crucial for stable demos and offline CI. • CSP without inline scripts forces better frontend design. • Documentation clarity significantly influences the evaluation experience for judges.
What's next for TutorFlow
• A modern React or HTMX front-end for smoother interactions • Mobile-optimized week calendar • Optional sync with Google Calendar / Apple Calendar • Exportable invoices (PDF templates) with customizable branding • Tutor-student communication features (messages, shared files) • Cloud-ready deployment template (e.g. Fly.io / Railway)
Log in or sign up for Devpost to join the conversation.