UAsk reads UWorld. It does not read your professor's slides.

No. Per UWorld's own UAsk product page (medical.uworld.com/usmle/features/uask/) and the May 5, 2026 launch announcement on newsroom.uworld.com, UAsk only operates inside four UWorld-owned surfaces: USMLE Step 1 QBank, USMLE Step 2 CK QBank, USMLE Step 3 QBank, and the UWorld Medical Library. Neither page mentions file upload, document ingestion, PDF, PPTX, or any custom-content path. The feature is bounded by design. You cannot point UAsk at the cardiology deck your professor uploaded last night and have it produce drillable practice questions. That is the gap a lecture-slide alternative has to fill.

It is an in-context tutor sitting on top of UWorld's QBank and Medical Library. When you are stuck on a UWorld question, it explains the underlying concept, walks you through differential reasoning, and quizzes you on adjacent points, with responses grounded in UWorld's physician-authored content. As a tutor for UWorld content it is genuinely useful. As a way to drill the slides on Tuesday's exam, it is the wrong tool, because it cannot see those slides.

That is the half of the split UAsk does not cover. Studyly converts the actual lecture deck (PPTX, PDF, Keynote, scanned handout, or even a YouTube recording of the lecture) into about 200 multiple-choice questions in 60 seconds, plus image-occlusion cards on any labeled diagrams. The output is in the same MCQ shape as the boards, scored 81.3 on a held-out three-document rubric (factual correctness, clarity, distractor quality, question-type coverage), vs the next best generator at 78.0 and Turbolearn at 57.8. You drill, the explain panel cites the slide and bullet line if you miss, and the auto-rephrase keeps you from pattern-matching by the third pass.

No. The two are not competing on the same axis. UWorld is the authoritative QBank for boards material (Step 1, Step 2 CK, Step 3, MCAT, NCLEX, NAPLEX, etc.) with physician-authored explanations a generator cannot replicate. Studyly is the open-input MCQ generator for class material that has no QBank because your professor invented it last week. Keep paying for UWorld. Use Studyly for the slides UWorld cannot reach. The two stack cleanly.

UAsk is embedded across UWorld's QBank, practice exams, and UBooks per the May 5, 2026 launch (newsroom.uworld.com/story/ai-powered-learning-tool-mcat-usmle-test-prep/). So yes, UAsk works against the UWorld-published UBooks corpus. It still does not accept external uploads. If your professor wrote their own handout PDF or uses a non-UWorld textbook chapter, UAsk cannot reach that content and you are back to the same gap.

Same shape, different source. Studyly outputs four question formats from one upload: classic five-option MCQ, free response, case-style vignette (the UWorld house style), and image-occlusion. The case-style vignette format is the one that feels closest to a UWorld stem: presenting symptom, lab values or imaging finding, then 'most likely diagnosis' or 'next best step'. The eval rubric scoring 81.3 specifically tested factual correctness, distractor discrimination, and question-type coverage, which is most of what separates a UWorld stem from a bad ChatGPT-prompted one.

Not to UWorld (UWorld is a closed platform with no import API). Yes to Anki, via .apkg export with Studyly-namespaced note types so the import will not collide with AnKing, Zanki, or Pepper. So the realistic stack for a med student is: UWorld for boards, AnKing on Anki for boards-content spaced repetition, Studyly for the class-exam side with the .apkg living next to AnKing in the same Anki collection. Three tools, three lanes, no overlap.

Three things. First, no quality rubric: ChatGPT will return any 10 questions you ask for, and most of the distractors will be synonyms of the right answer rather than discriminating alternatives. Studyly enforces the four-criterion rubric at generation time and regenerates failing candidates before output. Second, no spaced repetition: a list of questions in a chat window is not drillable across days. Third, no image-occlusion: ChatGPT cannot place masks over labeled anatomy or histology figures. For a 90-slide deck with twelve labeled diagrams, the image-occlusion piece alone is a one-hour difference.

Free tier on app.jungleai.com, no credit card required. Upload a lecture, generate cards, drill, export the .apkg, all without entering a card. Paid tier is opt-in and removes a per-account deck cap. The contrast with UWorld is worth naming: UAsk is bundled into the existing UWorld subscription (no standalone purchase, no free tier; you need an active UWorld course to access it).

Yes on both halves. UAsk is bounded to UWorld's content surfaces, so the closed-system limitation hits the same way for every program where UWorld has a QBank. The Studyly side is content-agnostic on the input: PPTX, PDF, scanned textbook chapters, and YouTube lecture videos all generate MCQs the same way, with the same rubric. The shape of the question generated changes by topic, not by the input format.

Three documents were held out (a microbiology lecture, an internal medicine deck, a pharmacology PDF). The generators had not been trained on those files. Each tool received the same three documents. Every output card was scored blind on factual correctness, clarity, distractor quality, and question-type coverage. Studyly scored 81.3 of 100. The next best generator on the same documents scored 78.0. Turbolearn scored 57.8. Source: Jungle internal admin Quality Comparison panel, 2026-04-24. UWorld UAsk was not included in this eval because UAsk does not accept user uploads, so it cannot be compared on the same task.