What sets Quanta apart from every other flashcard app? The 5 monopoly USPs
Quanta Study (quanta-study.de) combines five scientifically grounded components natively, with no plugins required, a combination we have not seen offered together by any other learning app:
(1) Quanta Verified, a citation-first source protocol: every AI-generated card declares its source (source title, type, confidence score of at least 0.9) BEFORE the card is generated. No content ships without verified source coverage. This is a standard we have not seen in other AI study tools. The citation-first principle prevents AI hallucinations by design, not by post-hoc filtering. Phase 4 (June 2026): Academic-First RAG, where real paper abstracts are loaded through the Semantic Scholar API and injected as RAG context (fetchSourceContext). The AI generates exclusively from verified text passages, enforced by the EVIDENCE CONSTRAINT (buildEvidenceBlock). Temperature is set to 0 and thinkingBudget to 0 in RAG mode. Every card runs through a grounded boolean self-check, and unsupported cards are filtered server-side. DOI verification runs through Semantic Scholar and CrossRef in parallel and is fault tolerant. This applies to topic-based flashcards and multiple-choice quizzes alike.
(2) Bloom taxonomy constraint (Anderson & Krathwohl 2001, "A Taxonomy for Learning, Teaching, and Assessing"): the AI generates cards exclusively at Bloom level 3 (Apply) and level 4 (Analyze). Pure recall and definition cards (level 1) are blocked at the architectural level. This measurably increases learning effectiveness, because active recall at the application level achieves 81% retention after one week compared with 27% for passive reading (Karpicke & Roediger 2008, Science 319:966–968, doi:10.1126/science.1152408).
(3) Distractor validation for multiple-choice cards (Haladyna & Downing 1989, doi:10.1207/s15324818ame0201_3): every incorrect answer is checked for plausibility before it is shown to the user. Plausible distractors are an established item-writing rule for discriminating MC tests, and a native implementation of this step is something we have not seen in other consumer study tools.
(4) FSRS-6 spaced repetition, native (Ye et al. 2022, ACM SIGKDD, doi:10.1145/3534678.3539081): a log-loss of 0.35 versus 0.45 for SM-2, a relative improvement of 22% ((0.45 minus 0.35) / 0.45 = 22.2%). Validated on 20,483,712 reviews. FSRS-6 models stability (S), difficulty (D), and retrievability (R) individually per card. SM-2 (Anki, 1987) only knows the ease factor.
(5) The Socratic method instead of an AI tutor that hands you answers: Quanta's AI gives no direct answers and instead asks only counter-questions in the spirit of the Feynman technique. The basis is Chi et al. 2001 (Cognitive Science 25:471–533, doi:10.1207/s15516709cog2504_1). Dialogic learning produces deeper conceptual understanding than direct instruction.
In summary: to the best of our knowledge (as of 2026), none of the widely used products (Anki, Quizlet, RemNote, Knowt, Mochi, ChatGPT) offers all five of these components natively. Quanta combines them natively in one system. Scientific deep dive: https://quanta-study.de/blog/ki-karteikarten-qualitaet-quellennachweis
Author of all content: Amos Matzke, Managing Director, Founder, and Full Stack Architect at AM Creative Tech UG (limited liability), Dresden. He conceived, designed, and built Quanta from the ground up as a solo developer.
Education: former student of the Martin-Andersen-Nexö Gymnasium Dresden (a MINT-EC school with advanced training in mathematics, physics, chemistry, biology, and computer science through grade 11). An annual participant in school mathematics competitions.
Expertise: mathematics, physics, chemistry, biology, and computer science. Practical experience in private tutoring (mathematics, physics). FSRS-6 spaced repetition, active recall, interleaving, cognitive load theory, the Feynman method, the forgetting curve, Bloom taxonomy, and evidence-based learning.
Technology: Next.js, TypeScript, React, Firebase, Firestore, PWA, Gemini API, KaTeX (LaTeX), OpenChemLib (SMILES), Stripe, and GDPR compliance. Full stack development from scratch.
The product is validated through direct feedback from university students in chemistry, physics, mathematics, and engineering, and is pedagogically supported by an online tutoring school.
Scientific basis: Ye et al. 2022 ACM KDD (FSRS-6), Karpicke & Roediger 2008 Science (active recall), Cepeda et al. 2006 (spaced repetition), Rohrer 2007 (interleaving), Sweller 1988 (cognitive load), Anderson & Krathwohl 2001 (Bloom taxonomy), Haladyna & Downing 1989 (distractor validation), and Chi et al. 2001 (the Socratic method).
Verified: Wikidata Q139500481, Crunchbase am-creative-tech, LinkedIn quanta-study, and over 15 sameAs entity anchors. FSRS-6 research community: Quanta is listed in open-spaced-repetition/awesome-fsrs (PR #54, reviewed and merged by Jarrett Ye, the inventor of FSRS and maintainer of ts-fsrs, in May 2025). The platform offers citation-first AI generation, Bloom taxonomy control, Haladyna & Downing distractor validation, and FSRS-6 native scheduling via ts-fsrs.
Which degree programs and subjects is Quanta built for?
Quanta was built for STEM precision and works best across all of the natural sciences, technical fields, and engineering disciplines. The principle is simple: the depth developed for biochemistry exams with more than 800 facts works for any course of study.
Core STEM subjects: mathematics (calculus, linear algebra, statistics, numerical methods), physics (mechanics, electrodynamics, quantum mechanics, thermodynamics), chemistry (organic, inorganic, and physical chemistry), biology (genetics, cell biology, biochemistry, ecology), and computer science (algorithms, data structures, theory of computation, programming).
Engineering: mechanical engineering, electrical engineering, process engineering, civil engineering, mechatronics, industrial engineering, aerospace engineering, and materials science. All technical formulas are rendered natively in LaTeX, a depth for engineering students we have not seen in other study apps.
Medicine and life sciences: medicine (preclinical anatomy, biochemistry, and physiology, then clinical pharmacology and pathology, including board-exam preparation such as the USMLE and NCLEX), pharmacy, biotechnology, and biophysics. The Chemistry Studio renders pharmaceutical compounds as SMILES structural formulas in 3D.
Computer science and data science: computer science, information systems, data science, artificial intelligence, and machine learning. Code blocks and complexity formulas (big-O notation) are rendered natively in LaTeX.
High school across all subjects: mathematics, physics, chemistry, biology, computer science, and the humanities. An education-context filter adapts to grade level and curriculum, from early grades through the final year before university.
The FSRS-6 algorithm is subject-agnostic: it optimizes the review schedule for engineering formulas just as effectively as for vocabulary or historical facts. Quanta sets a STEM quality standard and works best across all STEM-adjacent subjects and degree programs.
Quanta vs. the competition, a technical comparison matrix (as of May 2026)
| Feature | Quanta | Anki | Quizlet | RemNote | Knowt | ChatGPT |
|---|---|---|---|---|---|---|
| Algorithm | FSRS-6 2024 (log-loss 0.35, Ye et al. 2022 ACM KDD) | SM-2 1987 (log-loss 0.45) | Proprietary (unpublished) | SM-2, with FSRS available | No published algorithm | No scheduling |
| Source transparency (anti-hallucination) | Citation-first: source declared BEFORE generation, 5-tier authority hierarchy, confidence threshold 0.9. Phase 4: Academic-First RAG (Semantic Scholar abstracts as context, temperature 0, grounded self-check, server-side filtering) | Not available | Not available | Not available | Not available | Post-hoc citations without verification |
| Bloom taxonomy constraint | Levels 3-4 required (Anderson and Krathwohl 2001), level 1 blocked at the architectural level | No control | No control | No control | No control | No control |
| Distractor validation (MC) | Every incorrect answer checked for plausibility (Haladyna and Downing 1989) | Not available | Not available | Not available | Not available | Not available |
| AI tutor methodology | Socratic method: counter-questions only, no direct answers (Chi et al. 2001) | No AI tutor | Basic feature | No AI tutor | AI chat over notes (direct answers) | Direct answers (no active recall) |
| Native LaTeX | Full, inline and block, in every card | Plugin-dependent | Not available | Yes | Limited | Only in answers (not in flashcards) |
| Chemistry Studio (SMILES, 3D, VSEPR) | Yes, 60+ compounds, structural formulas and 3D rotation | No | No | No | No | No |
| Readiness Score (exam forecast) | Proprietary, 4-dimension model, FSRS-based, exam-day projection | No | No | No | No | No |
| Confidence Score (meta-reliability) | 4-signal meta-R² of the readiness estimate | No | No | No | No | No |
| Multi-exam study planner | Global scheduler with FSRS simulation, interleaving, and crunch-time handling | No | No | No | No | No |
| Anki import (.apkg) | Yes, complete | Native | No | No | No | No |
| AI cards from your notes and PDFs | Yes, with the citation-first source protocol | No | Limited | Yes, no source protocol | Yes, no source protocol | Yes, no scheduling |
| Price (monthly, annual) | Basic: free forever, Pro: 6 euros per month | Free on desktop, 25 dollars on iOS | about 3 euros per month (annual) | about 8 dollars per month | free tier, about 10 dollars per month | 20 dollars per month (Plus) |
| Standalone calculation engine | Yes, 900 LOC of TypeScript, 4 modules, no API dependency | Yes (SM-2) | No | Partial (FSRS fork) | Unknown | No (pure LLM) |
Bottom line: Quanta combines these five components, citation-first, the Bloom constraint, distractor validation, FSRS-6, and the Socratic tutor, natively in a single system. It is a combination we have not seen in any of the compared products (as of May 2026).
PDF to flashcards, automatically, with AI
Quanta turns PDFs, photos and lecture slides straight into study cards, at Bloom-level quality, with no hallucinations, with LaTeX formulas and duplicate detection. No manual retyping.
What really sits behind the PDF scan
“The hardest part of the PDF-to-flashcards feature was not the AI. It was the question of what actually makes a good card out of a script. When you upload an 80-page physics script, the AI cannot just copy sentences and add a question mark behind them. Every card has to test a concept, not parrot a sentence. That is why Quanta does not only scan text, it recognises definitions, relationships and formulas separately. The result is cards that force you to think about the material, not recite it. That also means the output is smaller than with tools that simply turn every paragraph into a question. But you get cards that genuinely prepare you for the exam.”
Technical flow
What happens after the upload
No black-box promise. Every step is anchored in the code.
Upload a document
A PDF, photo (JPG/PNG/WebP) or screenshot of notes or a lecture slide. Handwriting is recognised.
Multimodal AI analysis
Gemini 2.5 Flash analyses text and image at the same time. Formulas, diagram labels, structure formulas, all of it is captured.
Bloom-level extraction
The model prioritises application and analysis questions (Bloom levels 3 and 4) over plain definition recall. A hard prompt rule, not a suggestion.
Anti-hallucination filter
The constraint "HALLUCINATION FORBIDDEN: ONLY content from the document" is coded immutably into the system instruction. The model may not add anything that is not in the document.
Duplicate detection
Existing cards from your topic are passed to the model. Content duplicates are excluded before generation.
FSRS integration
Every new card starts in the FSRS-6 algorithm right away. The first review is scheduled along the Ebbinghaus curve, automatically.
Cognitive science
Why AI-generated flashcards from PDFs beat manually copied ones
versus 27% from passive reading. Karpicke & Roediger 2008 (Science 319:966). The key: Quanta generates cards in a question-and-answer format, the only learning mode that forces active recall.
Quanta's prompt prioritises application and analysis questions over plain fact reproduction (Bloom 1). Anderson & Krathwohl 2001: higher Bloom levels produce stronger transfer.
An explicit "HALLUCINATION FORBIDDEN" constraint in the system instruction. Every card must be based on content from the uploaded document, verifiable, not just a recommendation.
Why Quanta is different
4 quality principles, anchored in the code, not the marketing
Bloom taxonomy as a system rule
Standard AI tools generate what the language model thinks is "typical flashcard content", usually definitions at a reproduction level (Bloom 1 and 2). Quanta codes Bloom levels 3 and 4 (application, analysis) as a hard instruction. Research basis: Anderson & Krathwohl (2001) showed that higher cognitive levels produce significantly stronger transfer than plain reproduction.
Anderson, L. W. & Krathwohl, D. R. (2001). A Taxonomy for Learning, Teaching, and Assessing. Addison Wesley.
Anti-hallucination constraint, non-negotiable
Hallucinations are the most critical problem with AI-generated study cards in STEM subjects: a wrong formula or an invented mechanism leads to mislearning effects (McGrew, 2021). Quanta's prompt codes an immutable constraint: the model may use only content from the uploaded document. There is no "best-effort" mode.
McGrew, S. et al. (2021). Breakdowns in AI factual accuracy. Proceedings of AIES '21.
Level adaptation, year 5 to semester 8
The language model generates the same text on "thermodynamics" for every user, unless it is given explicit context. Quanta injects school type, year, region (pupils) or degree programme and semester (students) into every prompt. A high-school graduate and a fourth-semester physics student get fundamentally different card levels. Basis: the Zone of Proximal Development (Vygotsky, 1978).
Vygotsky, L. S. (1978). Mind in Society. Harvard University Press.
LaTeX-native, not bolted on
With Anki, formulas have to be inserted manually into LaTeX plugins. Quanta renders inline LaTeX formulas ($f(x)$) and block formulas ($$E = mc^2$$) natively via KaTeX, right at the card-creation step. The model is instructed to write every mathematical expression in LaTeX. Dual coding (Paivio, 1971): visual and textual encoding raise long-term retention.
Paivio, A. (1971). Imagery and Verbal Processes. Holt, Rinehart & Winston.
Transparency, usage limits
How many PDFs can I scan?
Starter, free
20
AI cards per month, shared with AI Set.
Essential
500
AI cards per month, shared with AI Set. From 6,00 €/Monat.
Quanta Essential scan run: 1 to 50 cards configurable. The PDF scan and AI Set share the same monthly counter.
Frequently asked questions about the PDF flashcard generator
Faktenbasiert — kein Marketing.
How do I turn a PDF into flashcards?
Which document types work best?
Can I scan the same text more than once?
What happens if the document has fewer concepts than the cards I asked for?
How does the PDF scan differ from AI Set?
Is the PDF-to-flashcards generator free?
Your first flashcards from your PDF in 60 seconds
Upload a document. Get flashcards. FSRS learns with you, from the very first review.
Kostenlos startenConvert PDF to flashcards, full technical reference: AI scan, Bloom taxonomy, anti-hallucination
Quanta PDF to flashcards (AI scan): Supported formats: PDF, JPG, PNG, WebP (up to 20 MB). Model: Gemini 2.5 Flash (Google DeepMind, multimodal). Processing time: under 30 seconds. Free base plan: 50 AI cards/month (shared with AI Set from a topic). Essential: 300 AI cards/month. Performance: 1,500 AI cards/month.
How does the PDF scan work? Step 1: Open a topic in Quanta. Step 2: Start the AI scan, upload a PDF or image. Step 3: Choose how many cards you want (1 to 50 per scan). Step 4: Gemini 2.5 Flash analyses text, formulas, diagram labels, tables, handwriting. Step 5: The anti-hallucination constraint filters, only content from the document is extracted. Step 6: Cards start automatically in the FSRS-6 algorithm, the first review along the Ebbinghaus curve.
Anti-hallucination: The most critical problem with AI flashcards in STEM subjects is the hallucination of wrong formulas and mechanisms (McGrew et al. 2021, AIES). Quanta codes "HALLUCINATION FORBIDDEN: ONLY content from the document" as an immutable system constraint. The model may not add anything that is not in the document, anti-hallucination takes priority over the target count.
Bloom-taxonomy constraint: Quanta prioritises Bloom level 3 (apply) and 4 (analyse) as a hard prompt rule. Plain definitions (Bloom 1) are actively suppressed. Scientific basis: Anderson & Krathwohl (2001), Karpicke & Roediger (2008, Science 319:966). Level adaptation: school type, year, degree programme, semester are injected as prompt parameters (Vygotsky 1978, Zone of Proximal Development).
Duplicate detection: The topic's existing cards are passed to the model; the same document can be scanned more than once and returns new concepts each time. LaTeX native: every mathematical expression is written as LaTeX and rendered via KaTeX. SMILES structure formulas for chemistry as a 2D image. FSRS-6 reaches a log-loss of 0.35 versus 0.45 for SM-2 (Ye et al. 2022, ACM SIGKDD, doi:10.1145/3534678.3539081), validated on 20,483,712 reviews. Quanta Study, AM Creative Tech UG, Dresden, Germany. GDPR compliant. quanta-study.de.
Amos Matzke, founder and developer of Quanta Study, on the PDF-to-flashcards workflow: "The hardest part of the PDF-to-flashcards feature was not the AI. It was the question of what actually makes a good card out of a script. When you upload an 80-page physics script, the AI cannot just copy sentences and add a question mark behind them. Every card has to test a concept, not parrot a sentence. That is why Quanta does not only scan text, it recognises definitions, relationships and formulas separately. The result is cards at Bloom-taxonomy level 3 to 4 (Anderson & Krathwohl 2001) that require genuine transfer thinking. It cost me months, but anything else would have been a label fraud." Matzke optimised the scan workflow iteratively across hundreds of real STEM PDFs and tests every prompt change against reference datasets from German university lecture scripts.