Flagship·evaluation

AI + Human Mains evaluation, on every answer you write

Submit a typed answer or a handwritten PDF — get a published-rubric score, dimension-level feedback, and a model-answer skeleton in seconds. Higher tiers add a vetted human evaluator inside 24 hours.

Try it nowSee plans
UPSC GS 1–4
10 Optionals
Essay paper
Handwritten PDF
India-first
gs2.v1.0 · Sonnet
Cooperative federalism
score
38/50
  • Content accuracy8/10
  • Structure7/10
  • Examples & evidence6/10
  • Analytical depth9/10
  • Conclusion8/10
Para 2 cites Article 263 but the answer skips Punchhi Commission's recommendations on inter-state council activation — add a single line on dispute-resolution architecture for full marks.
Live preview
<10s
AI pass on Haiku
5/5
rubric dimensions
24h
human-evaluator SLA

Why teams pick it

Three things you get on day one

  1. Rubric-aligned scoring

    Every answer is scored on the published 0–10 rubric across five dimensions, with a total out of 50 — no opinionated drive-by takes.

    Benefit 1 of 3

  2. Sentence-level rewrite tips

    Improvement roadmap references your actual paragraphs — "your para 2 lists X but skips Y" — not generic copy-paste advice.

    Benefit 2 of 3

  3. Human review on flagship plans

    Officer + Collector tiers route the answer to a vetted UPSC evaluator inside 24 hours, alongside the AI pass.

    Benefit 3 of 3

How it works

From input to output, in three

  1. Step 1input

    Pick subject + paper

    GS 1–4, Essay, or any of the ten Optional papers. Choose Mains or Optional sub-paper from the closed list.

  2. Step 2process

    Paste or upload

    Type your answer in the editor, or drop a handwritten PDF — Textract OCR handles scanned pages with handwriting on.

  3. Step 3output

    Get scored in seconds

    AI pass returns within ~10 seconds on Haiku tiers; structured JSON drives the dimension breakdown, roadmap, and model snippet.

Built for

Three readers we kept in mind

Aspirant
use case

Writes 4–6 mock answers a week and needs scoring tonight, not next week.

Working professional
use case

Has 90 minutes after work — needs structured, actionable feedback to course-correct fast.

Repeat candidate
use case

Knows where they fell short last attempt; wants tier-by-tier rubric scoring to track delta.

Under the hood

Capabilities that actually ship

  • Typed + handwritten PDF flows (S3 pre-signed upload, async Textract OCR)
  • Per-tier model selection — Haiku, Sonnet, Opus — driven from your plan
  • Prompt-cached system rubrics so repeat evaluations stay fast and cheap
  • PII redactor screens email/phone/Aadhaar before model invocation
  • Refusal handling for empty / off-topic / prompt-injection submissions
  • Single-active subscription invariant — quota cannot be double-spent

Related

The rest of the toolkit

FAQ

Questions worth answering

Ready when you are

Start with the free tier.No card needed.

Aspirant is free for 7 days, auto-granted on first Google sign-in. Upgrade only when you want human evaluator review or unlimited usage.

Get started — it’s freeSee pricing

Sign in with Google · we never see your card · cancel from settings.

  • On the free tier
  • 2 free AI evaluations
    Full rubric scoring on the Aspirant tier
  • All five tools
    Reader · Flashcards · PYQ · AI Tools · Eval
  • No card needed
    Auto-granted on first Google sign-in
  • 7-day window
    Upgrade only when you actually need more