Roadmap · Updated November 2026

One account for university life. Studying comes first.

Uni42 is four connected products for university students: a quiz app, an AI study suite, campus news and help getting into university. Study is the revenue engine; campus news is the moat. We build them in a fixed order, starting at the University of Calabria (Unical). Each phase has a clear goal, and the next one starts only when that goal is met.

Founder

Built from inside the campus it serves

Mourad Amri

Founder. Management and Finance student at the University of Calabria (Unical).

mouradamri.com
  • A student at Unical, the first campus: direct access to the associations that will publish on News and the students who will use Study.
  • Strong ties with Unical professors. One has already shown interest in developing the ready made courses.
  • Sales and customer background: energy consulting for private and business clients, and front line customer service.
  • National winner of FAI's La Tua Idea Green competition in 2022.
  • Directed the full design of Uni42 shown on this site: brand, every screen and this roadmap.
The plan

Four products, built in order

Each phase reuses what the one before it built, and earns before the next one starts.

  1. Phase 1Live 30 days after funding

    Uni42 Quiz

    Students add their own notes and Uni42 turns them into multiple choice quizzes. iPhone and web, one account, always in sync. First campus: Unical.

    • PDF, Word, text and pasted notes
    • Topics found automatically
    • Three difficulty levels
    • Weekly study goal and stats
    • English and Italian
    • Uni42 Study beta for trusted testers

    Why this order

    The smallest product with a clear job to be done. It builds the user base and the AI pipeline (reading documents, finding topics, writing questions) that every later product reuses, so each new product launches to an audience we already have.

    How it earns

    Free to drive adoption. Daily limits are enforced on our server, so the AI cost of every student is capped from day one and unit economics stay predictable.

    Goal to move on

    • v1 live on the App Store and the web, 30 days after funding
    • 1,000 sign-ups within 80 days of v1
    • 60% activation: first quiz finished in the first session
    • 30% week 4 retention
  2. Phase 2Summer exam session

    Uni42 Study

    An AI study suite built on the student's own material. Upload a PDF or notes once, and everything else works from it. Built in month 1 alongside the quiz app and tested with a trusted group; Phase 2 opens it to everyone.

    • A tutor that answers questions about your material
    • A live voice tutor on Gemini 3.8 Live and GPT-Live-1.1 that explains, then tests you with an oral exam
    • Quizzes and flashcards generated from your files
    • Ready made courses in maths, algebra, microeconomics and more, developed with university professors
    • A chart maker for subjects like economics, such as supply and demand curves

    Why this order

    Same users, same documents, much deeper engagement. This is where Uni42 starts recurring revenue, sold to an audience we already own.

    How it earns

    Freemium subscription. The free tier stays as the acquisition funnel; the voice tutor, oral exams, premium courses and higher limits are paid. Monthly and yearly plans.

    Goal to move on

    • Free to paid conversion on active users
    • Positive gross margin per paying student: revenue above AI and infrastructure cost
    • Monthly recurring revenue growing month over month
  3. Phase 3Next academic year

    Uni42 News

    Campus, course and class updates, written and published directly by student associations. Next to them, ads from local businesses around the university. Association partners are signed during Phase 1, so News opens with content on day one.

    • Posts from student associations
    • Updates by campus, course and class
    • Student accounts verified with a university email
    • Local business ads, clearly labelled

    Why this order

    This is the strategic core. Study is opened around exams; News is opened every day. Every association that joins brings its members, so distribution grows organically, and a campus where every association publishes on Uni42 is very hard for a competitor to take. Study is the engine that funds it.

    How it earns

    Local advertising. A verified student audience within walking distance of a shop is worth more than a generic social media ad, so each placement sells at a higher price.

    Goal to move on

    • Associations publishing every week
    • First paying local advertisers
    • Daily active students on News
  4. Phase 4Following enrolment season

    Uni42 Onboarding

    Help for people getting into university: choosing the right course, then step by step guides for documents, registration and the first weeks.

    • Help me choose a course
    • Guides for documents and registration
    • Tutorials for the first weeks on campus

    Why this order

    Top of the funnel. It reaches students before their first day, so they arrive already using Uni42. Onboarding acquires them, Study and News retain them.

    How it earns

    Partnerships with student services such as housing and transport, clearly labelled. And a direct path into Study subscriptions.

    Goal to move on

    • Incoming students using the guides before enrolment
    • Conversion from Onboarding into Study

Day one is the day funding arrives. Later phases are timed to the university calendar. Phase goals are the conditions to move on, not results achieved.

Why it compounds

Each product makes the next one cheaper to grow

  1. 01

    Onboarding brings students in before day one.

  2. 02

    Study keeps them through every exam session.

  3. 03

    News brings them back every day.

  4. 04

    Local businesses pay to reach them, which funds the next campus.

Business model

Three revenue streams on one user base

  • Uni42 Study

    Subscriptions

    Recurring revenue from students on monthly and yearly plans: the voice tutor, oral exams, premium courses and higher limits.

  • Uni42 News

    Local advertising

    Hyperlocal ads: businesses near campus pay a premium to reach a verified student audience that lives and studies around them.

  • Uni42 Onboarding

    Partnerships

    Student services pay for labelled placements in the guides new students follow.

Risks and how we handle them

Decided before we write a line of product code

AI cost grows with users
Daily limits counted on our server, and an AI gateway that lets us switch to cheaper models without rewriting code.
Wrong answers from AI
Questions are written only from the student's own material, so every answer can be checked against the source.
Student data and privacy
Encrypted storage, database rules that stop anyone reading another account, AI providers set to zero data retention, and a delete account button that erases everything.
NotebookLM, Quizlet and ChatGPT
They already make quizzes and flashcards from notes, and we don't compete on that alone. Study gets students in and pays the bills. What they don't have is the campus: associations publishing to their members, course and class updates, local businesses and the path into university. That network is built campus by campus and is hard to copy from outside.
AI agents building the product
The founder orchestrates them: sets each agent's goal, makes the product decisions and approves every release. A security agent and a review agent check every change, and nothing ships without human sign off, security review and legal review.
Building four products at once
We don't. Each phase has a goal, and the next one starts only when that goal is met.
Month 1 budget

30 days from design to a working product

Day one is the day the budget is in place. In the 30 days that follow, five AI agents working in parallel turn the finished design into Uni42 Quiz on iPhone and web, built, secured, tested and live as v1, and build the Uni42 Study beta alongside it for a trusted group of testers. From there, the goal is 1,000 students within 80 days. Every line is a dependency, and each one says what stops without it. Security and testing take nearly a third of the agent budget, because students trust us with their notes.

$8,000Month 1: design to working product
  1. AI agent team

    Estimate
    $6,200

    Persistent AI agents, each given one long-running responsibility instead of single tasks. Each one plans its own work, runs parallel sessions, keeps context across weeks, follows up on stuck work and opens pull requests. The founder orchestrates the team: sets each agent's goal, makes the product decisions and approves every release. They run on prepaid usage credits. In month 1 all five run in parallel, around the clock, so the whole Phase 1 build happens in 30 days instead of 6 months. The work is sized from screens and rules that are already designed.

    • Engineering agent: Uni42 Quiz v1, both apps and the server$2,000
      • Database for accounts, folders, documents, topics, quizzes, answers, weekly goals and daily limits, every table locked to its owner
      • Sign in with Apple and 6-digit email codes, on iPhone and on the web
      • File pipeline: PDF, Word and text upload, text extraction, private storage per account
      • AI pipeline: finding topics and writing questions in the background, with retries, failure states and live sync to both apps
      • Daily limits counted on the server: 5 uploads and 10 quizzes per account
      • Web app: sidebar, folders, drag and drop upload, quiz with keyboard shortcuts, stats, settings, delete account
      • iPhone app: 14 screen groups, each with its empty, loading, error and limit states, plus offline reading
      • Push notifications: daily reminder and quiz ready
      • Weekly goal and streak rules, stats over the last 8 weeks
      • Three transactional emails, and English and Italian throughout
      • App Store build and submission
    • Engineering agent: Uni42 Study beta, built in the same sprint$1,200
      • Every uploaded document indexed for search, so answers come from the student's own material
      • A tutor that answers questions about that material and shows where each answer comes from
      • A live voice tutor that explains a topic out loud, then runs an oral exam and scores it, built on the two newest real-time voice models: Google's Gemini 3.8 Live and OpenAI's GPT-Live-1.1
      • Both voice models behind one switch, tested side by side on quality, response time and cost per hour
      • Flashcards generated from the student's files
      • A chart maker for subjects like economics, such as supply and demand curves
      • Feature flags and staged rollout: every Study feature opens to a trusted tester group first, then to everyone in Phase 2
    • Security agent: always on, from the first line of code$1,200
      • Reviews every pull request for security issues before it can merge
      • Tests every database rule: no account can ever read another account's folders, files or quizzes
      • Checks uploads for file type, size and private storage access
      • Rate limits and abuse protection on sign-in codes and on every AI endpoint
      • Scans every build so no secret key ever ships inside the apps
      • Daily dependency audit, with fixes opened as pull requests
      • Verifies that delete account erases everything, every time
      • Keeps Study features locked to approved testers until each one passes security review
      • Alerts a human straight away on anything serious
    • QA agent: tests everything before students see it$800
      • Automated tests for every core flow: sign in, add a file, find topics, build a quiz, answer, see the result
      • Every screen checked in light and dark mode, in English and Italian
      • VoiceOver labels, Dynamic Type and keyboard use on the web
      • The same account checked on iPhone and web at once, so both stay in sync
      • Voice tutor sessions tested on both models for accuracy, response time and both languages
      • A full test run before every release, and follow-up on any work that gets stuck
    • Review agent: holds every change to the design standard$400
      • Every pull request checked against the design system: colours, type, spacing and corner radii
      • Every built screen compared with its mockup
      • Every line of copy checked against the brand voice and the product brief
      • Code quality and size rules enforced on every change
    • Growth agent: go-to-market for 1,000 students in 80 days$600
      • App Store optimisation in English and Italian: listing, screenshots and keywords
      • An always-on content engine for TikTok, Instagram and student group chats
      • Campus distribution through student associations: posters, QR codes and referral posts
      • A website page for each university we launch at
      • Weekly student interviews, turned into product changes and content
      • Partner agreements with Unical student associations, signed during the 80 days so News opens with content
      • A weekly funnel report: sign-ups, activation, week 4 retention and cost per student, against the 80 day goal

    Without it: No one builds, secures or launches the product. Uni42 stays a design.

  2. Infrastructure

    Public price
    $500

    Production grade from day one: the priority build tier, two seats, monitoring. Monthly list prices, plus the yearly Apple fee paid up front.

    • Supabase Pro: accounts, database, live sync, files$25
    • Vercel Pro, two seats: web app and API$40
    • Expo EAS Production: priority iPhone builds and updates$199
    • Resend Pro: sign-in codes and emails$20
    • Apple Developer Program, one year$99
    • Error monitoring, domain and support mailbox (estimate)$117

    Without it: There is no server, no database, no sign-in email and no App Store listing.

  3. AI usage

    Estimate
    $600

    The real AI calls behind testing. Quiz: about 10,000 test quizzes and 3,000 documents, at about 2 cents a quiz and 3 cents a document ($290). Voice tutor beta: about 60 hours on Gemini 3.8 Live and 50 hours on GPT-Live-1.1 (about $0.05 a minute) for the trusted testers ($250). Tutor chat and flashcards ($60). Daily limits cap the worst case on our server.

    Without it: Quizzes and the voice tutor can't be tested on real material before launch.

  4. Legal review

    Estimate
    $500

    A fixed-fee review of the privacy policy and terms by an EU lawyer, before the first student account is created.

    Without it: We can't legally store student accounts and files in the EU.

  5. Reserve

    Estimate
    $200

    About 2.5% for price changes, extra iPhone builds and a second round of App Store review.

    Without it: One rejected review or price change stops the launch.

Total, month 1$8,000

Public prices checked in November 2026 at each vendor's pricing page. Lines marked as estimates are our own calculation, shown with their working. Agent credits are sized from the designed scope.