Skip to main content

Technology Stack

🛠️ Technology Stack (Current Snapshot)​

The RTS prototype combines modern web, database, and AI technologies to scaffold student inquiry, reflection, and narrative development.

🧩 Frontend​

  • Framework: React 18 + Next.js (App Router)
  • Styling: Tailwind CSS
  • UI Components: shadcn/ui + Radix UI
  • Form Handling: React Hook Form + Zod
  • State/Server Sync: TanStack React Query
  • Authentication UI: Supabase Auth UI + Cookie Sessions

🗄️ Backend​

  • Database: Supabase (PostgreSQL) with Row Level Security (RLS)
  • Authentication: Supabase Auth (email/password + cookie-based session)
  • API Routes: Next.js API handlers (not yet using Edge Functions in production)
  • Storage: Supabase Buckets (planned for media, not actively used yet)

🤖 AI Integration​

  • Models Used:

    • Gemini 1.5 Pro + 2.5 Pro flavors, LearnLM (via Google Gen AI SDK)

  • Features Leveraged:

    • Multi-turn dialogue via stored Supabase thread messages
    • Structured output (e.g., JSON via Gemini’s responseSchema)
    • Function calling (tested, not active in all flows)
    • Text-to-speech (VolcEngine, tested)
    • Whisper (speech-to-text, tested)
    • Threaded reasoning via custom prompt construction
    • AI Transparency & Token Analytics (thinking summaries, token counts)
    • Structured JSON output via responseSchema + pure markdown synthesis
    • Course-based content management and assignment workflows

🌐 External APIs (Tested/Planned)​

  • Academic APIs: Semantic Scholar, Crossref, Ex Libris Primo, Google Scholar
  • Media APIs: Listen Notes (podcast search), Bing News
  • Speech/Audio: Whisper, VolcEngine TTS (tested)

🎯 Black Box Micro-Engagement (BBME) System​

  • Content Management: Template-based BBME creation with course assignment
  • 8-Part Reflection Structure: Action step + 7 reflection dimensions
  • Instructor Tools: Creation, assignment, and management dashboard
  • Student Workflow: Course-scoped BBME access and submission
  • AI Synthesis: Gemini-powered reflection analysis with full metadata capture

🧠 Architectural Philosophy​

  • Pedagogy-Oriented Orchestration:

    • System design scaffolds thinking, not just output
    • Conversations are not ephemeral — structured artifacts like reflections, keywords, and journal entries are persisted
    • Thread messages (from AI and user) are stored for synthesis and longitudinal analysis

  • Instructor-Student Ecosystem:

    • Walled garden course structure with RLS-enforced boundaries
    • Structured assignment workflows (Movement 1, BBMEs)
    • Complete audit trail of student work and AI interactions

  • Reflections, not Answers:

    • Every API route is a compositional aid, not a solution generator
    • System is designed for recursive, inquiry-driven use across movements