In developmentProcess-oriented AI maths teaching assistant
See the reasoning.Not just the result.
ShaktiMath is being built to read how a student solves a problem, locate the step that needs attention and turn that insight into the next useful learning action.
ANALYSIS TRACEPROBLEM 07 / 12
01 (x + 3)² = 3302 x² + 6x + 9 = 3303 x² + 6x = 33 − 904 x² + 6x = 24
STEP 03
The method is sound. Check the subtraction before continuing.
BENCHMARK PLATFORM · CHERRYPOT BY ZEZEDU
NanoSkool is using a proven international reference model while shaping an India-ready ShaktiMath. These figures belong to CherryPot and are not yet NanoSkool product results.
01 / WHY IT MATTERS
A correct answer can hide weak reasoning. A wrong answer can hide real understanding.
Conventional marking collapses a learner's entire thought process into a tick, a cross or a score. ShaktiMath is designed to make the path visible—so students know what to fix and teachers know where the class needs help.
Enough individual feedback
Immediate guidance on the exact step that needs another look, followed by practice at the right level.
Less grading, better intervention
Automate repetitive checking while keeping professional judgement, review and classroom decisions with the teacher.
Progress they can understand
Move the conversation from “What was the score?” to “Which idea is becoming secure?”
Insight beyond averages
See recurring misconceptions, monitor cohorts and direct support before gaps become exam outcomes.
02 / THE LEARNING LOOP
One piece of working becomes the next teaching decision.
Work naturally
Students solve on the web or app, write with a stylus, photograph a page or complete the paper assignment a school already uses.
PC · tablet · mobile · paperRead every step
Math-specialised AI recognises the written work, reconstructs the method and evaluates each stage against a generated rubric.
OCR · reasoning trace · partial creditGive useful feedback
Students see what was correct, what changed and how to improve. Teachers can review or adjust the result before sharing it.
Immediate · specific · teacher-reviewedPractise the right idea
The engine generates aligned variations, while dashboards reveal individual and class patterns for the next lesson.
Personalised practice · class insight03 / CORE CAPABILITIES
Built around mathematical process—not answer checking.
Each capability connects student action to teacher judgement rather than replacing either one.
Read real working
Math-specialised handwriting recognition turns stylus work, photographed pages and scanned assignments into analysable mathematical steps.
Benchmark OCR accuracy: 98%Find the broken step
The analysis follows the learner's method line by line, identifies where the reasoning changed direction and explains what to revisit.
Benchmark feedback accuracy: 97%Award partial credit
AI-generated rubrics recognise valid method, not only final answers. Teachers can review, edit and add their own feedback before results are released.
Teacher remains in controlGenerate the next problem
The maths engine can vary numbers, formulas, functions, diagrams and graphs while preserving the concept and reasoning being assessed.
Benchmark library: 300K+ problemsPersonalise practice
Follow-up questions are curated from each learner's progress so practice responds to the misconception, not merely the chapter.
Web · tablet · mobileSee the class clearly
Dashboards turn individual working into class-level insight, with exportable reports and planned integration into the systems schools already use.
Dashboard · report · LMS04 / TWO SCHOOL WORKFLOWS
Use it for daily learning. Use it without abandoning paper.
Personalised classroom practice
- PrepareSet up the class, syllabus and schedule; optionally upload teacher-owned material.
- AssignUse AI-curated practice or select the exact problems the class should solve.
- RespondDeliver process-level feedback automatically, with an optional teacher-led review session.
- UnderstandBring individual progress and high-level class insight into one dashboard.
Paper-based unit quizzes
- CreateUpload teacher content, generate key solutions and produce multiple exam versions.
- Run as usualStudents complete the paper-based assignment in the familiar classroom format.
- Scan and reviewUpload scripts for rubric-based grading, then let the teacher verify or change results.
- ActGenerate student and class reports, plus a focused preparation session with instant feedback.
05 / THE MATHS ENGINE
Same concept. Fresh route. Unlimited useful variation.
The benchmark engine can transform text, formulas, diagrams, functions and graphs. It can also analyse a teacher's own problem, preserve its structure and generate new variations with worked answers—supporting fairer assessment and faster preparation.
300K+ benchmark problems8K+ benchmark problem types
ORIGINAL(x + 3)² = 33Find x² + 6x
→VARIATION 01(x + 3)² = 18Find x² + 6x
VARIATION 02(x + 5)² = 41Find x² + 10x
VARIATION 03(x − 2)² = 29Find x² − 4x
06 / SCHOOL FIT
Designed to enter the system you already operate.
Internationally benchmarked
CherryPot currently supports IGCSE, IB, US and Korean National Curriculum contexts. NanoSkool's India-ready alignment will be confirmed as development progresses.
Connect the classroom stack
Reference integrations include Google Classroom, Microsoft Teams and commonly used LMS/SIS environments, with custom enterprise connections possible.
Teacher-reviewed by design
Teachers can select content, modify grades, add feedback and decide when a review session is needed. Automation handles repetition, not responsibility.
07 / SCHOOL PILOT
Help shape the India-ready ShaktiMath.
We are inviting schools to discuss a focused experience programme lasting from one month to one semester and covering up to three classes. The exact scenario and commercial proposal will be built around the school's curriculum, workflow and integration needs.
Request a pilot discussionCORE PILOT SCOPE
- Personalised curriculum
- Instant process-level feedback
- Common class assignments
- Partial-credit grading
- Dashboard and reports
Optional reference services include paper-material grading, problem generation, student/class analytics, preparation sessions, on-demand tutoring and university-admission consulting.
08 / COMMON QUESTIONS
Clear answers before a school commits.
Is ShaktiMath available now?
It is in development. NanoSkool is speaking with schools early so the India-ready product reflects real curriculum, assessment and integration needs.
Is this an answer-checking app?
No. The core proposition is process-level analysis: recognising mathematical working, evaluating each step, preserving partial credit and explaining where reasoning needs attention.
Does it require students to abandon paper?
No. The reference workflow supports web and app input as well as collected, scanned and uploaded paper assignments.
Can teachers change the AI result?
Yes. The reference model lets teachers review individual grading, modify results and add their own feedback before using the data for instruction.
Where do the performance figures on this page come from?
They come from the CherryPot introduction brochure supplied by ZEZEDU and describe its international benchmark deployment. They are clearly labelled here because NanoSkool's own pilot evidence has not yet been established.
BUILD WITH US