StudyMate
A student app where users upload materials and get AI-generated quizzes.
StudyMate, a weekend project, shows how a non-coding PM can ship real features with an AI workflow. Zevi Arnovitz added fill-in-the-blank questions (30% of tests, six answers for two blanks) from Linear ticket STU88, with the build taking minutes, and localized from Hebrew to English in two days. Lesson: chain explore, plan, execute, review and document steps, and capture ideas as tickets quickly.
- 1Dictate the specWispr Flow; 30% fill-in-the-blank tests
- 2Claude asks questionsA few short clarifying questions
- 3File Linear ticketTicket STU88 via MCP
- 4Explore and planScope questions, then markdown plan
- 5Cursor builds itComposer finishes in minutes
From Zevi Arnovitz's StudyMate workflow
What was said about StudyMate
StudyMate is a weekend side project used to demo a no-code AI workflow. Zevi Arnovitz, a Meta PM with no technical background, built StudyMate so students can upload their materials and get quizzes generated by Gemini. It only offered multiple choice questions, while competitors offered true or false and fill in the blank.
He runs it as a company of one with all the context in his head, which is why he files tickets more loosely than he would at work. The workflow was shown on this solo project with no teammates. No bug rates, costs or time comparisons are given.
A dictated idea became a shipped fill-in-the-blank feature within minutes. He dictated the spec with Wispr Flow: 30% of tests as fill in the blank, six answer options for two blanks, drag and drop. Claude asked a few short questions and filed it in Linear via MCP as ticket STU88. He then explored the scope, wrote a markdown plan, and had Cursor's Composer build it in minutes.
The ticket held a TLDR, current state, expected outcomes and context. Exploration had Claude ask about scope, data model, UX, validation, grading and prompt changes. He says a human engineer would need days or a week for the same work, though the live build skipped the full exploration answers, review and real testing.
Cursor showing the update-docs.md slash command, which tells Claude Code to update documentation after code changes. Parallel AI agents localized StudyMate from Hebrew to English in two days. In one week he ran three agents in parallel: the StudyMate Hebrew-to-English localization took two days, alongside a personal site live on a domain in 90 minutes and podcast prep.
He calls these "time machine moments." He does not track what the AI credits cost and counts them as tuition for learning.
How a Meta PM ships products without ever writing code | Zevi Arnovitz
A weekend project used for the live build. It only had multiple choice, so he added fill-in-the-blank questions (30% of tests, six answers for two blanks, drag and drop) via a Linear ticket STU88, plan and Composer build in minutes. He also localized it from Hebrew to English in two days.
StudyMate is a weekend side project used to demo a no-code AI workflow. Zevi Arnovitz, a Meta PM with no technical background, built StudyMate so students can upload their materials and get quizzes generated by Gemini. It only offered multiple choice questions, while competitors offered true or false and fill in the blank.
He runs it as a company of one with all the context in his head, which is why he files tickets more loosely than he would at work. The workflow was shown on this solo project with no teammates. No bug rates, costs or time comparisons are given.
A dictated idea became a shipped fill-in-the-blank feature within minutes. He dictated the spec with Wispr Flow: 30% of tests as fill in the blank, six answer options for two blanks, drag and drop. Claude asked a few short questions and filed it in Linear via MCP as ticket STU88. He then explored the scope, wrote a markdown plan, and had Cursor's Composer build it in minutes.
The ticket held a TLDR, current state, expected outcomes and context. Exploration had Claude ask about scope, data model, UX, validation, grading and prompt changes. He says a human engineer would need days or a week for the same work, though the live build skipped the full exploration answers, review and real testing.
Cursor showing the update-docs.md slash command, which tells Claude Code to update documentation after code changes. Parallel AI agents localized StudyMate from Hebrew to English in two days. In one week he ran three agents in parallel: the StudyMate Hebrew-to-English localization took two days, alongside a personal site live on a domain in 90 minutes and podcast prep.
He calls these "time machine moments." He does not track what the AI credits cost and counts them as tuition for learning.
More in Education
EducationDuolingo
A language learning app built around daily streaks and short lessons.
47 sources · 132 tips
EducationHeadway
A book summary and self-growth app that sells subscriptions through a long web quiz funnel.
10 sources · 36 tips
EducationBlinkist
A nonfiction book summary app that turns books into short audio and text reads.
16 sources · 17 tips
EducationStudyFetch
An AI study platform that turns course materials into practice tests, flashcards and tutoring sessions.
4 sources · 31 tips
EducationCoconote
An AI study app that records and transcribes lectures and creates notes, summaries and flashcards.
3 sources · 30 tips
EducationReforge
A company offering courses on product and growth, now also building AI products like Reforge Insights.
10 sources · 13 tips
EducationCoursiv
An AI learning app that teaches beginners practical AI tools through short lessons and challenges.
2 sources · 17 tips
EducationUnacademy
An Indian online learning platform offering courses for exam preparation and other subjects.
2 sources · 6 tips