Exact course identity
Qualification + board + subject/course + specification code + exam year are kept together, preventing a generic course label from silently controlling the AI.
VERSION-AWARE CURRICULUM INTELLIGENCE
The AI is given the student's actual qualification, board, course, level and exam-year context.
XYSHIMA uses structured curriculum records rather than treating labels such as “Physics” or “Mathematics” as universal courses. The system can distinguish qualification, awarding body, specification code, SL/HL or course variant, exam year, topic structure and assessment context before a diagnostic, quiz or tutoring session is generated.
SPECIFICATION ENGINE
Structured facts, not copied specification documents.
XYSHIMA stores concise curriculum data such as official identifiers, topic and skill taxonomies, course level, assessment structure and version notes. It does not reproduce awarding-body specification prose. Each built-in record also carries a coverage status so the AI knows whether the map has been verified as structured coverage or whether an uploaded official syllabus should be treated as the authoritative source.
Version-aware
Where awarding bodies publish syllabus windows, XYSHIMA can resolve the version from the student's examination year rather than assuming one permanent syllabus.
Verified vs broad-scope status
The tutor is explicitly told when a structured map is verified and when it is only a broad board-level map. It is instructed not to claim exact coverage where verification is incomplete.
Official syllabus override
A student can upload the current official specification or syllabus. Retrieved material from that document becomes authoritative for fine-grained course grounding.
Board-aware pathways are built into course selection.
The curriculum layer supports the principal boards used by XYSHIMA's current school-subject families and is designed for ongoing specification verification.
GCSE
AQA, Pearson Edexcel, OCR, Eduqas, WJEC and CCEA are represented in the board layer. Verified subject records include official codes where they have been checked; remaining records stay explicitly marked as broad-scope until their official specification is verified.
IGCSE / International GCSE
Cambridge International and Pearson Edexcel International course identifiers are supported for the current science, mathematics and economics subject families.
AS & A Level
AQA, Pearson Edexcel, OCR, Eduqas, WJEC, CCEA and Cambridge International can be selected. Cambridge syllabus windows are exam-year aware where known.
International A Level
Pearson Edexcel International modular unit identifiers are represented so the AI can distinguish an International Advanced Level pathway from a UK linear A Level.
IB · AP · SAT SUITE
Course variants are separated instead of collapsed into generic labels.
IB Diploma
Physics, Chemistry, Biology and Economics are available as SL and HL variants. Mathematics is separated into Analysis & Approaches and Applications & Interpretation, each at SL and HL.
AP
AP Physics 1, Physics 2, Physics C: Mechanics, Physics C: Electricity & Magnetism, Calculus AB, Calculus BC, Statistics, Chemistry, Biology, Microeconomics and Macroeconomics are represented as distinct courses.
SAT Suite
SAT, PSAT/NMSQT and PSAT 10 use the College Board Reading & Writing and Math domain/skill structure, while PSAT 8/9 has its own lower-level Math scope where the framework differs.
The curriculum map is injected into every AI learning request.
Before the OpenAI model generates a diagnostic, explanation, quiz or flashcard set, XYSHIMA adds a curriculum-control block containing the qualification, board, course code, version, exam year, structured examinable map, assessment notes and confidence status. A verified map tells the model to stay inside that scope; a non-exhaustive map tells it to avoid claiming exact coverage and prefer an uploaded official specification when precision matters.
Diagnostics stay course-specific
Questions are generated against the selected course rather than a generic subject. Topic and misconception metadata then feed the student's learning model.
Future updates are isolated
Curriculum records live in a dedicated specification layer, so a revised syllabus can be updated without changing the tutoring model or rewriting the rest of the application.
LANGUAGES
Curriculum grounding and teaching language are separate.
Students can choose English, Spanish, French, German, Italian, Portuguese, Greek, Arabic, Persian, Hindi, Tamil, Simplified Chinese, Traditional Chinese, Japanese or Malay while retaining the selected qualification's official terminology, notation and course identifiers.