Your CAT mentor: understand what went wrong, try a focused fix, and check whether it helps.
💬
Chat with Marg like a mentor — ask anything about CAT prep
📝
Daily RC, DILR and QA practice targeting your specific mistakes
📊
Track your progress and see your cognitive patterns evolve
📝
Daily Practice Tab
Choose RC, DILR or QA. Practise a topic, take a timed check, or work on a pattern you’re investigating.
How it works
Your attempts help Marg build a clearer picture over time. One wrong answer is a clue, not a label. Review the reasoning, then test the change on another exercise.
Ask Marg about any question
After answering, tap "Ask Marg to explain this" on any question to get a detailed breakdown in chat.
🎯
Chat Like a Mentor
Marg remembers you across sessions and builds on every conversation.
"Explain why I keep getting RC inference questions wrong"
"Analyse my mock — VARC 28, DILR 18, QA 34"
"I have 2 hours today, what should I focus on?"
"Walk me through this DILR set I got stuck on"
Before you go 👋
Quick feedback helps us improve Marg for you.
What could be better?
Free CAT workspace · No card
Studying for CAT, but your score isn’t moving?
Start from the evidence.
A score is not a cause, and a shared target is not a shared plan. Marg asks for the attempts, accuracy and time that decide the next step, and it says when a source has not been checked.
Not another study plan. Not another mock. A cause waits until a check can separate it from the other explanations.
Or open a specific workspace
Your destination
Where does it happen?
The pattern from the ad is first
20-second decision check
What this choice points to
Test this properly with Marg →
Choose honestly.
●Free CAT workspace·No payment details
Try Marg before signing in
One real RC decision. No account needed.
The answer matters less than the reason your option looked attractive.
Policies are often defended because their outcomes are measurable. Yet measurability can create an illusion of objectivity: what is easy to count begins to stand in for what matters. This does not make measurement useless. It means a metric should be judged partly by the decisions it hides, not only by the comparisons it enables.
Which option best captures the author’s central claim?
What Marg does
Everything a serious CAT aspirant needs.
💬
AI Chat Mentor
Ask anything — mock analysis, section tips, study plan, RC doubt. Marg remembers you across every session.
🎯
Daily Targeted Practice
Fresh RC, DILR and QA checks with a stated reason. When Marg has enough evidence, the next check tests that working pattern.
🧠
Cognitive Pattern Tracking
Marg separates what happened from what it may mean—then tests whether option traps, set selection or execution actually repeat.
📊
Mock Analysis
Paste your mock scores. Marg records where the marks fell, then asks for attempts, accuracy and time before it will name a cause.
📈
Progress Tracking
See your streak, session count, and cognitive patterns evolving over time. Know exactly how far you've come.
🔄
Session Memory
Marg remembers what you worked on last time. Every session builds on the previous one. The longer you use it, the better it knows you.
Use Marg because
One wrong answer is an observation. A cause has to survive a check.
The same missed question can be a concept, an execution miss, or a decision to stay too long. Marg keeps those apart, and a later correct item is not called mastery.
How it works
Evidence first. Then one next action.
1
Start from the baseline you actually have
Target, time, section scores, attempts, accuracy and minutes. If one of those is missing, Marg asks for that piece and does not invent a plan.
2
Separate the explanations
An untimed miss, a timed miss, and staying too long are different next exercises. The reply says which explanation is still untested.
3
Use a source only with its status
Official CAT, a named book, a teacher’s question, and AI-generated practice stay distinct. An unknown channel is not judged, and it is not called safe.
✓
Evidence can build across sessions
Marg keeps relevant attempts and outcomes, so later advice can use what was actually observed instead of restarting from zero.
FROM THE FOUNDER
"I'm Garv — I built Marg while giving CAT 2026 myself. I gave it twice before without proper guidance and felt the same frustration both times — not a lack of resources, but a lack of direction. Marg is what I wish existed."
— Garv Singhal, Founder · CAT 2026 aspirant
Evidence
before diagnosis
Checked
keys only, when a ledger says so
Free
always for CAT 2026
What counts as improvement
A later comparable attempt, not a quote
Marg reports whether attempts, accuracy and time moved after one named intervention. No movement is a result. A single mock is not a percentile.
— Measurement rule, not a student testimonial
Stop guessing. Start fixing the right thing.
Free. The first useful step is the missing evidence, not a timetable.
Google sign-in gives us your name, email and profile photo. Across sessions we may store chats, study profile, cognitive and behavioural patterns, diagnoses, mentor tasks, attempts and results, mock history, practice progress, check-ins, and product-engagement milestones. Question images may be stored privately with extracted text, section/topic and numbered-question metadata so you can refer to them later. If you allow notifications, we store the browser subscription, twice-daily setting and delivery records; browser permission stays under your control. A community-invite number is stored only when submitted, in Supabase—not Gemini or chat. Friend challenges store the question and anonymous open, answer, help, reshare and signup events; the creator stays private.
How we use it
Your data is used solely to provide and improve Marg's mentoring experience. Reported response text enters quality tests only when you keep the separate review permission selected; feedback still sends if you decline.
Data storage
Your data is stored securely on Supabase servers. We do not sell, share, or trade your personal information with any third parties.
Deleting your data
Individual question images can be deleted from Question Library, and deleting a chat removes its linked images and extracted questions. To request deletion of your account and all associated data, including stored images, email support@trymarg.com from the email linked to your Marg account. We will process the request within 7 business days. Clearing a chat or browser storage alone does not delete records stored in Supabase; use Question Library or chat deletion for those linked records, or contact us for account-wide deletion.
Contact
support@trymarg.com
Terms of Service
Last updated: May 2026
Acceptance
By using Marg, you agree to these terms.
What Marg is
Marg is an AI-powered CAT exam mentoring tool. It provides guidance based on information you share. Not a substitute for professional coaching.
Contact
support@trymarg.com
About Marg
Built in 2026
The story
Marg was built by Garv Singhal, a CAT aspirant who gave the exam twice and felt the same frustration both times — not a lack of resources, but a lack of direction.
The name
Marg means "path" in Hindi — your path, personalised and specific.
The longer you use Marg, the better it knows your mistakes.
Marg is your free AI mentor for CAT — it builds a cognitive profile of how YOU specifically think and generates daily practice targeting your exact weak patterns.
How Marg works
DAY 1
You tell Marg your mock scores and weak areas
EVERY DAY
Marg uses targeted practice to test a working read before calling it a pattern
EVERY WEEK
Marg shows exactly what improved and what to fix next
Let's set up your profile 🎯
30 seconds. Marg uses this to personalise everything.
Category
Academic Profile (%)
Current Mock Percentile
Weakest Section
Email (for streak reminders)
We'll remind you if you miss a day. No spam ever.
Step 1 of 4 — Your mock scores
What did you score?
Enter your raw scores from today's mock. Leave blank if you didn't attempt a section.
VARC
Max 72
DILR
Max 60
QA
Max 60
Step 2 of 4 — Your diagnosis
Here's what I'm seeing.
Based on your scores — here's the pattern that's costing you marks.
Analysing your mock scores...
Step 3 of 4 — Quick context
One quick question.
This helps Marg give you a more specific plan.
Is this your first CAT attempt?
Step 4 of 4 — Last one
Which section feels most broken?
Not just the lowest score — the one that frustrates you most.
✓ Profile analysed
Your IIM Risk Profile
Target percentile: —
⚠ Biggest threat to your IIM call
—
--:--
Did you study yesterday?
How many hours did you study?
2h
Slide to set hours
Mentor Home
Here’s what we should work on today.
Your CAT mentor: understand what’s going wrong, test a small fix, and see whether it helps.
Marg recommends
Not sure why your score is stuck?
Tell Marg what happened in your last mock. We’ll unpack one section at a time, then choose what to try.
Or do something else
Your context stays connected
Your personal CAT mentor
?
Loading streak...
You have 5 free messages left — login to save this conversation
Your profile and progress stay connected. This conversation starts fresh.
📖 Today’s VARC · based on Aeon
Train the part of RC that is costing you marks.
One original CAT passage · 4 questions · stored key, not an independent check
What usually goes wrong when you attempt an RC?
What do you want to practise?
Difficulty
Reading world
Balanced diagnosis · CAT-level · surprise theme
Marg will choose a fresh original theme after you start.
Photos ready to send
Up to 4 pages
Press Enter to send · Shift+Enter for new line
Mock Analysis
Scores are the starting evidence—not the diagnosis.
Understand what happened in your mockEnter your VARC, DILR and QA scores, or upload a scorecard screenshot. Then tell Marg about your attempts, time and difficult moments. Scores alone cannot tell us the cause.For example: “I spent 22 minutes on my first DILR set and couldn’t finish another.” Marg helps separate a difficult set from a decision about when to leave it.
Your latest mock scores
Previous Analyses
Open original chat
No mock analyses yet.
Sectional Tests
Timed interfaces with navigation, submission and analysis.
Pressure-test one decision at a timeChoose a VARC reading world, focused QA topic, or DILR family. Each card states how its answer key was checked.
VARC
Timed RC sectional
1 embedded passage · 4 questions · timer and review. The stored key has not been independently checked.
QA
Focused QA sectional
10 solver-checked QA questions · 20 minutes · one selected topic. Local numeric solver, not an independent review.
DILR
DILR sectional
12 solver-checked DILR questions · 40 minutes. Local solver only. No published DILR bank is in this pilot.
Daily Practice
Tap an option to submit your answer. Chat messages do not select an option in Practice.
Marg is generating today's personalised practice based on your profile...
Your Progress
Updated after each session
-
Sessions
-
Mocks tracked
-
Day streak
Your Marg journey
Pattern, evidence, test and next step.
Building evidence
Your evidence trail appears after a diagnosis and targeted check.
Your challenge trail
Successful means a friend actually answered—not merely opened the link.
0
Friends reached
0
Friends opened
0
Friends answered
Mock Score History
VARC
DILR
QA
Your Cognitive Profile
VARC Pattern
Marg is learning your VARC patterns — keep practising
DILR Pattern
Marg is learning your DILR patterns — keep practising
QA Pattern
Marg is learning your QA patterns — keep practising
🎯 Marg's Weekly Assessment
📊
Progress tracking starts here
Give your first mock analysis and start daily practice. Marg will track your improvement every week.
🧠
Question Library
Your saved pages and the questions Marg found in them.
