Apar Exam
AI-Powered Examination & Assessment Platform
AparExam is a self-serve exam portal for independent educators. Pick a board, grade, subject, and chapters — the platform assembles a curriculum-aligned paper from a 300K+ question bank in under 2 seconds, exports a copy-protected branded PDF, and processes uploaded answer sheets through a five-step AI grading pipeline with OCR, per-question evaluation, and educator override. Generate and grade on credits: 5 free every month to try the full workflow before paying anything.
Select your syllabus. Generate the paper. Grade the answers. Done.
Most exam tools solve exactly one problem and stop there. A template editor gives you a blank page. A question-bank browser lets you browse but not assemble. A generic AI prompt box produces exam-looking text once and forgets the context the moment the session ends. An institution-first ERP handles the paper logistics but requires an admin, a contract, and a training cycle before a single educator can use it. None of them connect the full workflow from syllabus selection to completed graded submission. AparExam takes a different position. It is a self-serve educator product that compresses six connected steps — register, select curriculum, generate paper, export protected PDF, upload answer sheets, and review AI grades — into one credit-metered portal. The intelligence underneath is not rebuilt from scratch inside a consumer product. AparExam runs on top of the same paper assembly engine, question bank, and grading tools that power institution-facing products. That architectural choice is commercially meaningful. The portal gets real depth without rebuilding anything from scratch, and individual educators get a tool that behaves like serious exam infrastructure instead of a novelty generator. The commercial model matches educator behavior. Teachers work in bursts around exams, coaching schedules, and academic cycles — not on steady enterprise SaaS timelines. Paying per generation and per grading submission, with a persistent credit balance that does not expire, fits how individual educators actually consume a tool like this.
The complete paper generation workflow — select course, pick units, choose template, download paper and answer keys
Every screenshot shows the same four-step educator workflow: select the course and curriculum scope, pick chapters, choose a template, download the paper and answer keys. What the screenshots do not show is the assembly engine running underneath — drawing questions from the bank, balancing difficulty, enforcing template structure, applying copy protection, and deducting a credit — all completing before the page finishes rendering.
Exam tools that solve one step leave educators doing the rest by hand
The individual educator market is consistently underserved by exam tooling. Institutional products are built for admins, not for a single tuition teacher working alone. Lightweight tools help create content but abandon the educator at the moment grading begins. The gap between 'I have a paper' and 'I have graded results' is still measured in hours of manual work for most educators outside large institutions.
Most tools stop at content creation — the workflow does not continue
A paper generator that produces a Word document and nothing else leaves the educator to handle printing, distributing, collecting, grading, recording, and returning results entirely outside the tool. That means the tool saved perhaps 20 minutes of paper creation but did nothing about the 4 hours of manual grading that follow it. Workflow continuity is not a feature — it is the product.
AparExam connects paper creation, export, sharing, answer-sheet submission, and AI grading inside one credit account.
Generic AI generators forget the curriculum context
An AI prompt box that produces 'exam-style questions on Chapter 5 of Class 10 Science' once is not a repeatable tool. It has no question bank to draw from, no Bloom's taxonomy enforcement, no template structure, and no concept of the board's syllabus. Educators who try these tools once get an output. They cannot rely on the output being consistent, balanced, or aligned with what their students are actually being examined on.
AparExam's paper assembly uses structured curriculum drill-down: board → grade → subject → chapters. The output is grounded in an actual question bank, not free-text generation.
Manual grading is the invisible cost that no one builds for
An individual tuition teacher handling three batches may grade 60-90 answer sheets per exam cycle. Each sheet requires reading handwriting, applying partial marks, writing feedback, and tallying totals. At 8-12 minutes per sheet, that is 8-18 hours of work per exam cycle — before any teaching happens. Reducing that cost by even 70% changes the economics of running a small coaching operation.
AI grading with OCR, batch evaluation, and per-question override is the second product inside AparExam — not an afterthought feature.
Institution-first tools are too heavy for direct educator adoption
Most serious exam platforms require an institution account, an admin setup, a teacher profile under that institution, and often a contract negotiation before anyone can generate a paper. A tuition teacher working independently cannot adopt these tools because the product model assumes an institutional buyer. AparExam uses direct educator onboarding with simple ownership rules: if you created it, you control it. No admin hierarchy. No institution approval.
The portal deliberately does not use institution-scoped row-level security. Direct ownership is the product model.
Three educators, three distinct reasons to use a self-serve exam platform
AparExam is built for educators who work directly with students outside a large institution's IT infrastructure — or alongside one without depending on it. The product value is different for each, but the common thread is: they need a working tool today, not a procurement cycle.
Primary user
Independent tuition teacher or private coaching instructor
Runs one or more student batches independently. Creates their own exam papers, grades their own answer sheets, and manages their own schedule. Does not have access to institutional question banks, paper-generation tools, or grading infrastructure. Currently does everything by hand or from personal saved templates. Time spent on exam logistics directly competes with teaching time.
Small operation
Coaching institute operator with 3-10 faculty
Runs a structured coaching program with multiple subjects and regular exam cycles. Needs consistent paper quality across teachers, professional-looking output with the institute's branding, and a repeatable grading workflow that does not bottleneck on one person's availability. Institutional LMS platforms are too expensive and too complex for the scale. Free tools are too thin.
Dual user
School or B2B platform faculty who also teaches independently
Already uses an institution-facing platform through their school or employer. Also conducts personal tuition or workshops outside that institutional context. Needs a separate tool for the work they do independently — one that does not require switching accounts or re-entering identity. The B2B-to-dual upgrade path means one underlying account covers both contexts without duplication.
Six layers that turn curriculum selection into a graded result
AparExam is not a standalone product built from scratch. It runs on top of the paper-assembly engine, question bank, and grading tools that power institution-facing products — then adds the layers a self-serve product actually needs: direct onboarding, credits, branding, sharing, and human-review controls. Each layer below is a distinct decision about what the portal needs to own versus what it can borrow from shared infrastructure.
Built on shared infrastructure, not rebuilt from scratch
AparExam does not reimplement paper assembly or grading logic. It runs on top of the same question bank, paper assembly engine, and grading tools that power institution-facing products. On top of that shared foundation, the portal adds the layers a self-serve product needs: individual educator accounts, credit tracking, institutional branding on exports, paper sharing, and human-review controls for graded results. This is an architecture decision, not a development shortcut.
A self-serve product that gets its depth from shared infrastructure can focus its own effort on the things that make consumer adoption work: onboarding, credits, branding, and simplicity.
Curriculum selection drives everything — no free-text prompts
The paper generation flow starts with structured curriculum navigation: choose a board, then a grade, then a subject, then chapters. Educators think in syllabus slices and chapters, not in AI prompts. The selection flow matches how educators actually approach paper creation, which is why the output is grounded in the actual curriculum rather than an approximation of it.
'Chapter 5 of Class 10 CBSE Science' is a precise lookup against the question bank, not a semantic guess from an open text field.
Paper assembly: select → build → score → charge → deliver
When an educator submits their curriculum scope, the platform validates the selection, draws questions from the bank across the chosen chapters and template constraints, assembles the paper, scores its quality, deducts one credit, and returns the full paper — all in under 2 seconds. Saved preferences allow one-click reuse of the same configuration. Regeneration is free: the educator can request a different variation of the same paper without spending another credit.
The paper includes both AI-generated and textbook questions, supports optional or-questions, and carries a quality score. Credit consumption is the only irreversible step, and only for new generation — not for iteration.
PDF export with copy protection, branding tiers, and shareable links
The platform produces two export paths: a question paper (with optional answer space) and an answer key. Free-tier PDFs carry AparExam watermarks. Paid-tier PDFs include the educator's institution name in the header and footer, plus the institution logo. All exports are password-protected and locked against copying — including the accessibility copy permission that some PDF readers treat as equivalent to full copy access. Papers can also be shared via a secure link: enable sharing, get a URL, and recipients see the paper without any of the educator's account or billing details.
Institutional branding on the PDF is the most immediately visible upgrade value in paid plans — it changes the professional identity of the document, not just a backend balance number.
Five-step AI grading pipeline with OCR, per-question evaluation, and educator override
A grading submission is linked to the paper that was generated — so the question text, expected answers, total marks, and structure are already known before grading begins. The grading process runs five steps: load the paper's ground truth, extract text from the uploaded answer sheet images using OCR (with a separate process for Hindi scripts, which require different handling than English), match answer text to questions, evaluate all questions in a single AI call, and finalise the results. For multi-page answer sheets, each page is processed individually and combined. The system tracks progress, token usage, and estimated cost per grading job. Each question result stores the student's extracted answer, the expected answer, the marks awarded, AI feedback, and a confidence score — with fields for the educator to override both marks and feedback.
Grading all questions in one AI call reduces latency and cost. OCR handling for Hindi answer sheets is especially important given how common they are in Indian classrooms, where OCR reliability varies significantly across scripts.
Credit economy: simple rules, permanent upgrades, and a complete billing trail
The credit rules are simple: 1 credit to generate a paper, 1 credit to grade a submission, 1 credit to regrade, 0 credits to regenerate or export. Free users receive 5 credits every month. Paid credits never expire. Template access limits are set to the highest plan ever purchased and never decrease — so buying a higher plan is a permanent capability upgrade, not a subscription that resets. Every credit debit and refill is logged permanently in an audit trail that cannot be altered. Payments go through Razorpay, with signature verification and duplicate-rejection so a successful payment is processed exactly once.
Credits that do not expire and template limits that never decrease make the purchase decision feel like a permanent capability upgrade, not a time-limited subscription event.
Four flows that trace an educator from registration to graded results
These are the end-to-end sequences that connect an educator's intent to a working exam workflow. Each one is a design decision about what the platform should do automatically, where the educator retains control, and what the credit system should charge for versus treat as free iteration.
Educator registration and onboarding — two paths, one account system
Register an educator as a portal user, create their credit balance, and capture working context — board, grade, subject preferences — so the portal behaves like an educator workspace from the first session rather than a blank slate.
Why it matters: The dual upgrade path is a platform signal. It means Aparsoft does not treat the school product and the self-serve portal as unrelated silos. A school teacher who already uses an institutional platform can adopt AparExam without creating a second identity or re-verifying anything. Onboarding is not decorative profile setup — it allows the portal to remember preferred boards, grades, and subjects so the educator returns to familiar selections rather than rebuilding context every session.
Outcome: An educator registers once and lands in a workspace that already knows their preferred board, grade, and subject. Existing B2B users get portal access in one step without account duplication. Credit balance starts immediately — 5 free credits available before any payment decision.
Paper generation — curriculum drill-down to copy-protected PDF
Walk an educator from selecting their curriculum scope to downloading a professionally branded, copy-protected exam paper and answer key — with assembly completing in under 2 seconds and the credit deducted only once per generation act, not per iteration.
Why it matters: The template is the educator's only structural decision after picking the curriculum scope — it defines difficulty distribution, marks allocation, question type mix, and optional or-question structure. The assembly engine handles question selection, difficulty balance, and quality scoring without the educator touching those constraints. Regeneration consuming zero credits is a meaningful commercial choice: the platform charges for the generation act, then allows unlimited iteration on the same paper configuration. Sharing via a secure link gives students, parents, or co-teachers read-only access to the paper without needing a portal account.
Outcome: Educator selects board, grade, subject, chapters, and a template. Paper assembles in under 2 seconds. PDF exports with institutional branding (if on a paid plan) and copy protection. Paper is shareable via a secure URL. Educator can regenerate a different variation at no additional credit cost.
AI grading pipeline — upload to reviewed results
Process uploaded answer-sheet images or scans through a five-step AI pipeline — reading handwriting, matching answers to questions, evaluating each one against the known paper structure — and return per-question marks, feedback, and confidence scores the educator can inspect and override before finalising.
Why it matters: The pipeline starts from a known ground truth — the paper that was generated already contains the questions, expected answers, total marks, and structure. This is stronger than grading in isolation from an unknown paper. Hindi answer sheets require a different text-extraction process than English ones, and the system handles this automatically. Evaluating all questions in a single AI call — rather than one call per question — reduces both processing time and cost. If text extraction fails completely, the grading stops before attempting evaluation — no partial, unreliable results are silently produced. Credits are not refunded on grading failure because the AI processing cost has already been incurred — this boundary is documented, not hidden.
Outcome: Educator uploads answer sheets (single or multi-page). Grading completes asynchronously with progress tracking. Results arrive as per-question marks, feedback, and confidence scores. Educator reviews, overrides any individual question mark or feedback they disagree with, records remarks, and finalises. The machine does the first pass. The educator is the authority.
Saved preferences and quick-generate — turning repeated work into one click
Capture the educator's most-used paper configuration as a saved preference so recurring exam cycles — the same board, grade, subject, chapters, template, and difficulty — require one action instead of navigating the full generation flow every time.
Why it matters: A saved preference is a quiet but important product structure. An educator who runs the same class weekly or monthly — the same board, grade, subject, and chapters with the same template — should not reconstruct that setup every time. Quick-generate converts a saved configuration back into a generation run with one action. The same quality, the same difficulty balance, the same template constraints — without renavigating the curriculum selection each time. This aligns the tool with habitual educator behavior rather than one-time novelty usage.
Outcome: Educator saves their preferred configuration once. From that point, paper generation for recurring exam cycles is a single action. The full pipeline — question-bank retrieval, template enforcement, quality scoring, credit deduction — runs identically every time. The educator gets consistent output without repetitive setup.
What makes the sub-2-second claim credible and the credit system trustworthy
Speed and commercial integrity are both required. A paper generator that assembles fast but charges for stale outputs, or a grading system that fails silently without clear boundaries, is not a product educators can rely on. These are the implementation choices that make both claims hold.
Quality gates
Two simultaneous generation requests from the same account cannot both pass a 'credits available' check — the credit deduction is concurrency-safe by design.
Every credit debit and refill is recorded in a permanent audit log with its reason, amount, and timestamp — balance questions and disputes have a full trail to check.
Template access limits are set to the highest plan ever purchased and never decrease — a plan upgrade is a permanent capability gain, not a subscription that resets.
Payment signature verification runs on every Razorpay webhook — requests without a verified signature are rejected before any credits are granted.
Duplicate payment webhooks have no effect — each checks whether the credit grant has already been applied before executing.
On regeneration, previously downloaded PDFs are cleared — educators cannot accidentally download a stale version after requesting a new paper.
PDF copy protection covers all exports and specifically blocks the accessibility copy permission that some PDF readers treat as equivalent to full copy access — this is intentional, not an oversight.
If text extraction from answer sheets fails completely, grading stops before evaluation — no partial or unreliable results are silently produced.
Paper sharing links expose only a read-only view — the educator's account details, credit balance, and generation data are excluded from the shared output.
Portal access rules — requiring a completed educator profile and available credits — are enforced at the server level, not just in the interface. Direct API calls cannot bypass them.
Six things an educator can measure after using AparExam
These are the direct results of the architectural choices above — not aspirational claims. They are grounded in how the platform works, not in what the marketing says it does.
An educator selects their curriculum scope and template, submits the generation request, and receives a complete exam paper in under 30 seconds end-to-end. The assembly layer itself completes in under 2 seconds — the remainder is API overhead and PDF generation.
AI grading processes an entire answer sheet submission — OCR, answer segmentation, per-question evaluation — asynchronously. An educator reviews and overrides rather than grades from scratch. A class of 30 papers that previously took 6 hours takes under 2 hours with AI-assisted first-pass grading.
Five real monthly credits let an educator generate papers and grade submissions before making any payment decision. The free tier is not a demo — it runs the same pipeline as paid plans.
Paid-plan exports carry the educator's institution name in the header, footer, and logo position. Copy protection is applied to all exports — free and paid. The professional identity of the output changes with the plan, not just the backend balance.
Saved configurations and quick-generate reduce recurring exam cycles — same course, same chapters, same template — to a single action. The full quality-scored paper assembles identically every time without navigating curriculum selection again.
AI produces the first-pass grade — marks awarded, written feedback, and confidence score per question. The educator inspects each question result, overrides marks or feedback where they disagree, adds reviewer remarks, and finalises. The machine assists. The educator decides.
The advantage is a connected workflow on real shared infrastructure — not a single clever feature
A competitor can build a paper generator. They can add an AI grading endpoint. The harder part is the combination: a curriculum-grounded question bank feeding a sub-2-second assembly engine, copy-protected branded PDFs, a five-step grading pipeline with separate handling for Hindi answer sheets and per-question educator override, a credit system with permanent upgrade benefits, and seamless access for educators already on an institution platform — all in one portal an educator can onboard into in five minutes.
Depth comes from shared infrastructure, not from reinvention
AparExam's paper quality comes from the same question bank and assembly service that powers institution-facing products. An individual educator using the portal gets access to a 300K+ question bank, Bloom's taxonomy coverage, internal choice support, and a quality score — without any of that being rebuilt for a B2C product from scratch. A competitor building a consumer exam tool from nothing cannot match that depth on launch.
Credit model fits educator behavior — it does not fight it
Educators work in exam-cycle bursts, not on steady monthly subscription timelines. Charging per generation and per grading submission, with credits that never expire and free regeneration, matches how teachers actually use this kind of tool. An annual subscription would price out most individual educators. A per-use model where credits persist aligns commercial incentives with actual usage patterns.
Grading keeps the human in the loop at the right level
The grading system does not claim to be the final examiner. It produces a first pass: extracted text, matched answers, marks awarded, feedback, and a confidence score per question. The educator reviews and overrides. This is a more credible and commercially durable design than a system that overclaims accuracy and fails educators in practice. Per-question results show the student's answer, the expected answer, max marks, awarded marks, AI feedback, and override fields — built for inspection, not for blind trust.
Dual identity removes the friction of entering a new B2C product
An educator who already uses an Aparsoft institutional product — through their school or employer — can adopt AparExam without a new account, new login, or new verification step. A single upgrade converts an existing institution user into an exam portal user in one action. That removes the most common reason educators try and then abandon self-serve tools: the friction of managing parallel accounts for related products.
AparExam shows Aparsoft can consumerize deeper infrastructure into a self-serve educator product
Apar AI LMS proves Aparsoft can build institution-facing academic infrastructure. AparAcademy proves Aparsoft can build a direct-to-student learning product on top of shared educational intelligence. AparExam proves a third distinct capability: Aparsoft can take the same underlying exam engine and wrap it in a self-serve B2C product that an individual educator can adopt, use, and pay for without an institutional intermediary.
Usage-based monetization designed for real educator behavior
Credit-based billing, a real free tier, non-expiring paid credits, free regeneration, and durable template limits after plan upgrade — these are not default SaaS billing choices. They are deliberate decisions that match how individual educators consume a tool with uneven, burst-pattern usage. Aparsoft has built and operates this model in production.
Platform reuse over product duplication
AparExam does not rebuild the paper assembly or grading engine inside the portal. It runs on shared infrastructure and adds the self-serve layers — onboarding, credits, branding, sharing, and educator-controlled review — that make the product work for individual adoption. Knowing what to own versus what to borrow from the platform is a consistent engineering decision across all Aparsoft products.
Staged AI workflows where a single AI call is insufficient
The grading pipeline runs five explicit steps — not a single AI call and not a vague chain of prompts. Each step is distinct: load the paper structure, extract text, match answers, evaluate, finalise. The pipeline runs in the background with progress tracking, and if any step fails, it stops cleanly rather than silently producing unreliable partial results. Aparsoft applies multi-step AI orchestration where staged processing with conditional routing is the right approach — not as a default, but where the problem demands it.
Multi-product identity coherence — institution and self-serve in one account
The unified account upgrade path is not a feature added afterward. It is an early decision to keep both product surfaces on the same underlying identity layer instead of building separate user tables for each. That decision makes cross-product adoption frictionless and keeps billing, verification, and identity management unified. Aparsoft makes these decisions before they become expensive to fix.
