STARK Hackathon 2026 · Team XOR

Tell IFA your goal.
Get a course built around it.

IFA runs a three-model pipeline: a conversational advisor learns what you are trying to achieve, a course architect drafts a blueprint you approve, and a builder writes the lessons — grounded in ScholarXiv academic research and YouTube, generated one module at a time. Then run the whole app hands-free with Voxide voice navigation.

One-click demo learner on the sign-in page — no setup, no credit card, real data.

3-model pipeline ScholarXiv research Voxide voice JIT modules
IFA Studio · live pipeline live
I want to prepare for my exit exam — C# backend APIs.
Got it. How many hours a week can you realistically study?
About 6 hours.
Perfect — any material I should build from?
Model 1 is assembling your learner profile

Real integrations — not mock data

ScholarXiv Academic research
Voxide Voice navigation
Gemini Conversational AI
Groq Fast inference
YouTube Video sources
PostgreSQL Learner state

How it works

Three models, three jobs, one course that fits you

Instead of asking a single prompt to invent a curriculum, IFA splits the work. Each stage hands a clean, inspectable artifact to the next one — and you sit in the middle of it.

01 Model 1

Conversational advisor

You chat. IFA asks the questions that actually matter and turns your answers into a learner profile instead of guessing.

  • Understands your goal in plain language — no course codes
  • Captures hours per week, preferred creators, existing materials
  • Only proposes a course once your profile is complete
02 Model 2

Course architect

A reviewable blueprint lands before a single lesson is written — modules, key topics and hour estimates you can edit.

  • Module-by-module plan with topics and estimated hours
  • Paste links, docs or repos the course must consider
  • Nothing is generated until you approve the blueprint
03 Model 3

Course builder

Lessons and exam-style quizzes are written one module at a time, grounded in research and shaped by your assessment results.

  • ScholarXiv papers, web and YouTube gathered per module
  • Lesson + quiz generated on demand, saved to the course
  • Adaptive: weak areas return with more depth and practice
Chat your goal Review the blueprint Sources are researched Lessons & quizzes built

Platform

Everything you need after the course exists

IFA is a learning companion, not a content dump: courses you can actually enrol in, skills you can measure, research you can revisit, and peers you can compare against.

Three-model pipeline

Intake, architecture and building are separate models with separate jobs — not one prompt asked to do everything.

ScholarXiv research

Academic search through the ScholarXiv MCP, plus web and YouTube, saved as a reusable research package per course.

Voxide voice control

Talk to the app: jump to a section, start a course, or open the tutor. Voice maps to real capabilities, not a mock transcript.

Adaptive engine

Assessment outcomes feed back into the plan, so the next module leans into the skills you have not locked in yet.

Just-in-time generation

Modules materialize when you reach them. Course one opens with a finished lesson and quiz, not a placeholder.

Skills & progress

Skill breakdowns and progress analytics computed from your real assessment data, with weak areas called out.

Social learning

Share a course with an enrollment code, publish it to the marketplace, and benchmark your skills against peers.

Keyboard-first search

A ⌘K command palette jumps across courses, skills, research and pages without touching the mouse.

Voice navigation

Hands-free, powered by Voxide

Voxide maps your speech onto capabilities the app registers — so a spoken request runs the same code path as a click. No fake transcripts, no dead ends: if IFA cannot do it, it tells you.

Navigate

“Take me to my progress”

Moves the app to any section: home, courses, marketplace, recommendations, skills, progress, research, settings or tutor.

Create a course

“I want to learn GraphQL”

Runs the real course creation flow with your topic and weekly hours, then opens the generated course.

Open the tutor

“Explain pointers”

Opens the AI tutor focused on the topic you named, ready to explain or drill you.

Key detected A Voxide publishable key is configured, so the mic is armed after sign-in.
IFA voice console listening
Create a course on data structures
On it — 5 hours a week. Opening your new course…
Take me to my progress
Done. Here is your progress dashboard.

The browser SDK only ever uses the publishable vox_pub_… key; the secret vox_sk_… key never leaves the server. Whitelist your origin in the Voxide dashboard.

Research package ScholarXiv MCP
Academic papers Search run through the ScholarXiv MCP server ranked by relevance
Web resources Official documentation and reference pages fetched per topic
YouTube walkthroughs Matched to the creator you asked for transcript-ready

In the lesson

Sources are folded into the prompt and appended to the lesson under Course Sources — and your own links go in there too, so nothing you supplied gets dropped.

Research

Backed by ScholarXiv, not vibes

Research runs once per course and is stored as a reusable package, so every module you generate later is grounded in the same evidence — without paying to search again.

  • ScholarXiv academic search for peer-reviewed starting points.
  • Web references pulled for the specific topics in each module.
  • Video resources matched to the creator you prefer.
  • Your pasted links and cover image are treated as first-class sources.

3

AI models in the pipeline

9

learning surfaces in-app

5

domains supported

JIT

module-by-module generation

Built for exam prep across five domains

Computer science, engineering, medicine and health, science and business — the pipeline adapts to the subject, not the other way around.

Engineering Medicine & Health Computer Science Science Business

Under the hood

Engineered, not assembled

A typed SvelteKit client on top of a clean-architecture .NET API, with AI providers and research wired in as first-class services.

Frontend

  • SvelteKit 2 · Svelte 5
  • TypeScript
  • Tailwind CSS design system
  • Lucide icon set

Backend

  • ASP.NET Core (.NET 10)
  • Clean Architecture layers
  • Entity Framework Core
  • PostgreSQL

AI & integrations

  • Gemini — conversational AI
  • Groq — fast inference
  • ScholarXiv MCP — research
  • Voxide — voice interface
JWT auth & learner-scoped data • Responsive, theme-aware UI • Docker & Docker Compose
Free to try

Your next course is one conversation away

Describe the goal, approve the blueprint, and start learning from a research-backed module today. Or sign in with the demo learner and let Voxide drive.

IFA

Your AI learning companion. Personalized courses, research-backed lessons and hands-free navigation.

Study•Grow•Achieve

Explore

Get started

Built by Team XOR

  • Nathnel Teklemariam
  • Ermiyas Eshetu
  • Negede Tekleyes

STARK Official Hackathon 2026

© 2026 IFA. All rights reserved.

Research by ScholarXiv · Voice by Voxide