AI • Search • Marketing intelligence for Indian brands

विमर्श vimarsha: Sanskrit for careful reflection and examination

See what AI sees.

Vimarsh helps Indian brands understand how AI systems mention, recommend, and cite them across AI search -- with the exact response behind every number.

Get your free AI reportSee how it works

One free report on all four AI engines. No card needed.

  • 01

    Four AI engines

    OpenAI, Claude, Gemini and Perplexity, depending on your plan

  • 02

    Evidence for every number

    Open the raw response and its sources

  • 03

    Search Console, connected

    See which pages earn search clicks, beside your AI results

  • 04

    Built for India

    Questions localized to Indian cities

  • 05

    Eight kinds of question

    Brand, price, comparison, near-me and more

  • 06

    Weekly measurement

    Runs every Monday on the engines your plan includes

  • 07

    AI Visibility score

    One headline number for how AI treats your brand

  • 08

    Mention and recommendation

    How often AI names you, and how often it recommends you

  • 09

    Citation rate

    How often an answer points to a source

  • 10

    Sample size, always shown

    Every score says how many AI calls it rests on

  • 11

    Your competitors, named

    Comparison questions include the competitors you add

  • 12

    Ad connectors, connected

    Google Ads and ChatGPT Ads, alongside your AI results; Meta Ads rolling out

  • 13

    Billed in rupees

    Monthly plans in INR, through Razorpay

  • 14

    Delete on request

    Remove your account, with 14 days to change your mind

  • 15

    Breached-password check

    Signup rejects passwords known to have leaked

  • 16

    Local Radar Grid

    Scan AI visibility locality by locality, for physical businesses

  • 17

    Agency Intelligence

    One portfolio view across every client business you manage

  • 18

    Action Center

    Visibility gaps turned into prioritized, specific actions

  • 19

    Change Alerts

    Know when AI's description of your brand meaningfully shifts

Vimarsh is a product by Free Mind Consultancy

Vimarsh tells you your score. Our agency fixes it.

A done-for-you team that diagnoses why AI engines overlook you, fixes the causes, and re-measures to prove what changed -- built for how AI search actually works today. Alongside Vimarsh, Free Mind Consultancy also helps businesses with AI automation and consulting, SEO/AEO/GEO and AI search visibility, performance marketing, and publishing services.

See Agency plans

A quiet shift

More and more questions now go to AI. Is your brand in the answer?

ChatGPT, weekly users

In millions, at the dates OpenAI announced them.

02505007501000100MNov '23200MAug '24400MFeb '25800MOct '25900MFeb '26
ChatGPT weekly users in millions
Nov '23100
Aug '24200
Feb '25400
Oct '25800
Feb '26900

Questions asked, each day

Googleabout 13.7 billion
ChatGPTat least 2.5 billion

ChatGPT alone already handles roughly 18 questions for every 100 typed into Google.

And in India

0crore

people use ChatGPT every week (100 million), OpenAI's second-largest market after the United States.

While you read this, roughly searches have been made on Google and at least questions asked on ChatGPT.Every second, on average, about 1,58,549 searches are made on Google and at least 28,935 questions are asked on ChatGPT.

When a buyer asks an AI who to trust, it names a few brands and leaves the rest out. Vimarsh shows you whether yours is named, how, and why.

Sources: growth milestones as announced by OpenAI (Nov 2023 to Feb 2026); Google's stated “more than 5 trillion searches a year” (2025); Sam Altman on 2.5 billion ChatGPT prompts a day (July 2025) and on India (February 2026). The running line above is an estimate, not a live feed: a stated figure divided into seconds. A search and a prompt are different things; the comparison is one of scale, not like for like.

Built on the best, in the open

The engines are the refrigeration. Your brand is what people ask for by name.

Refrigeration made the cold drink possible; the brand is why people reached for it. AI works the same way. OpenAI, Claude, Gemini and Perplexity are the infrastructure your customers already ask, and Vimarsh does not pretend to be one of them.

We use each provider's official API, openly, to ask the questions your buyers ask and keep every answer. Our work is the layer above the engines: measuring, comparing and explaining what they say about your brand, so you can see where you stand and work to earn your name in their answers.

Which engines you get depends on your plan. OpenAI, Claude and Gemini are asked without web browsing, so their answers show what each model itself says about you.

Official APIs only
Each engine is asked through its own documented API. Nothing is scraped.
The same question, every engine
One buyer question goes to each engine your plan includes, worded identically.
Every answer is kept
The raw response is stored word for word, so any number can be traced back to it.

Vimarsh is independent. It is not affiliated with, sponsored by or endorsed by OpenAI, Anthropic, Google or Perplexity. Product and company names belong to their owners.

Through Vimarsh, let us see: are Indian brands visible in the ocean of AI models?

The gap

Your search ranking is not the whole story.

A page can rank first on Google and still be left out of the answer a customer gets from an AI engine. Pick a question and see how differently the engines place the same brand.

  1. IGoogle tells you where you rank.
  2. IIAI shapes the answer people hear.
  3. IIIVimarsh shows you what AI actually says about you.

Fig. 1 · Where each engine places your brand

Illustrative example

Choose a buyer question

Asked of every engine: “best project management software for small teams”

1st2nd3rd4thNot named
Your Google rankfrom your own SEO tool

0 of 4

engines name your brand first

3 of 4

name it anywhere in the answer

Based on 4 AI responses across 4 providers, this week

OpenAI · named first: Competitor A · your brand 3rd

“For a small team I'd start with Competitor A, then Competitor B. Your brand is worth a look if you want a simple setup.”

Invented figures with generic names. Vimarsh measures the engine rows; the Google marker is the rank you already know from your own SEO tool, shown for contrast.

Run on infrastructure you can trust

Built on infrastructure the world already relies on.

Vimarsh runs on widely used, proven infrastructure: the network that protects it, the database that keeps your evidence, the payments provider that bills you in rupees and the containers it ships in.

  • Cloudflare

    Edge protection

  • PostgreSQL

    Database

  • Razorpay

    Billing in rupees

  • Docker

    One container image per commit

  • GitHub

    Code, CI and image registry

Logos are trademarks of their owners and show technology Vimarsh uses. Their use does not imply endorsement or partnership.

How it works

Ask. Analyze. Explain.

What actually happens when you run Vimarsh, from one buyer question to the exact words behind every number.

Fig. 2 · From a question to its evidence

Illustrative example

A buyer question

“best project management software for small teams in Pune”

Eight kinds of question

BrandRecommendationProblem-basedNear mePrice inquiryComparisonAlternativesCategory discovery

For businesses in India, Vimarsh also adds city-localized and Hinglish versions of these questions.

Invented figures with generic names. Today you open a response on your dashboard to read its raw answer; the position table and the highlighted mention illustrate what the stored measurements contain.

The output

Your AI visibility, in one place.

After each run Vimarsh shows where you stand, with the response behind every number one click away.

Fig. 3 · Your visibility at a glance

Illustrative example
0100

0%

AI Visibility

Low agreement

Based on 16 AI responses across 4 providers, this week

Mention rate75%
Recommendation rate50%
Citation rate19%

By engine

0%25%50%75%100%75%OpenAI75%Claude50%Gemini100%Perplexity

Perplexity names your brand in 4 of 4 responses.

AI Visibility, last 6 weeks

38% 52%

Invented figures from 16 example responses. Today the dashboard shows your latest scores and each response behind them; the dial, the by-engine bars and the 6-week trend illustrate what the stored measurements can show. Only Perplexity searches the web, so it is the only engine that shows sources.

Competitor intelligence

Why does AI recommend them instead of you?

Competitors you name appear in the comparison questions Vimarsh asks the engines. Every response is stored and its mentions extracted, and competitor share is computed from them; no page shows that share yet.

Illustrative example · generic names

Illustrates stored measurements; the dashboard has no competitor view today.

Based on 48 AI responses across 4 providers, this month

How often the engine actually recommends each brand. Share of 40 recommendations across the set.

All shares, as a table (counts in brackets)
BrandMention shareRecommendation share
Your brand25% (15)20% (8)
Competitor A40% (24)45% (18)
Competitor B25% (15)25% (10)
Competitor C10% (6)10% (4)
Total100% (60)100% (40)

Competitor A was named first for “best project management software for small teams” in 3 of 4 responses.

Your brand was not named first in any of the 4 responses.

Who is named first, query by query

Five of the twelve example queries.

  • best project management software for small teamsCompetitor A · first in 3 of 4 responses
  • project management software with the best supportYour brand · first in 3 of 4 responses
  • most affordable project management softwareCompetitor B · first in 3 of 4 responses
  • project management software alternativesYour brand · first in 2 of 4 responses
  • project management software for startupsCompetitor C · first in 2 of 4 responses

Across all 12 queries: Competitor A named first most often in 7, Your brand in 2, Competitor B in 2, Competitor C in 1.

Citation intelligence

Which sites is AI actually pulling from?

When an engine answers with sources, Vimarsh records every one, so you can see which sites appear in the answers about your category -- and whether yours is among them.

A citation is a site named inside an AI answer. It is not a backlink.

Sources come from Perplexity only.

Of the four engines, only Perplexity searches the live web and returns the sites it used. OpenAI, Claude and Gemini are asked without web search, so they name no sources. Everything on citations below therefore comes from Perplexity's answers.

  • Perplexity
    Live web search. Returns the sources it used.
  • OpenAI
    Asked without web search. No sources returned.
  • Claude
    Asked without web search. No sources returned.
  • Gemini
    Asked without web search. No sources returned.

Illustrative example · generic names

Today the run page lists the sources cited in each response; this tally and the annotated answer illustrate stored citations.

Sources named across the Perplexity answers

Based on 12 AI responses across 1 provider, this month

50 citations across the 12 Perplexity responses. Citation rate: 12 of 48 responses across all four engines = 25%, because the other three name no sources.

  1. [1]industry-reviews.example14 · 28%
  2. comparison-blog.example10 · 20%
  3. [3]community-forum.example8 · 16%
  4. news-publication.example6 · 12%
  5. competitor-a.example5 · 10%
  6. [2]your-website.example (you)3 · 6%
  7. Other sources (each under 5%)4 · 8%

Illustrative example

One answer, with its sources

Perplexity · project management software with the best support · 2026-09-21
“Teams that prioritise support often shortlist Your brand first, followed by Competitor B . Reviewers point to response times, and one community thread compares onboarding help .”
  1. [1] industry-reviews.example
  2. [2] your-website.example
  3. [3] community-forum.example

Choose a [number] to see that source in the tally. A citation shows what the AI pointed to. It doesn't prove the source supports the claim.

Technology

An intelligence layer, not another model.

Vimarsh doesn't train, host or resell an AI model. It measures the ones your customers already ask -- and keeps the evidence. Here is how it's built, in plain terms.

  1. 1

    Your questions

    Buyer-style queries for your business

  2. 2

    Each engine's own API

    Official APIs, no wrapper, no scraping

  3. 3

    Raw answer stored

    Word for word, kept 90 days

  4. 4

    Mentions extracted

    Each with the span of text it came from

  5. 5

    Scores

    Every number links back to its answer

Talks to each engine directly

One adapter per engine -- OpenAI, Anthropic's Claude, Google Gemini and Perplexity -- each calling that provider's own API, with no third-party wrapper in between. The adapters share one interface, so every engine comes back in the same shape.

Nothing to install

No script, tag or SDK goes on your site, and Vimarsh doesn't need access to your code or analytics. It asks the engines from the outside, the way a customer would.

Keeps the source of every number

The raw response from every engine is stored, and each extracted mention records the span of text it came from. Raw responses are purged after 90 days; the extracted results stay.

Separates each customer's data

Every query is scoped to your organization in application code. Postgres row-level security adds a second layer on the routes that read or write your business data, so a query that forgets its scope still can't read another organization's rows there.

Bills through Razorpay

Subscriptions run on Razorpay. Its webhooks are accepted only with a valid HMAC-SHA256 signature over the raw request body, and each event is recorded once by its event ID.

Guards accounts

Passwords are hashed with bcrypt, must be at least 12 characters, and are checked against known-breached passwords at signup. Sessions live in httpOnly cookies, and the admin console requires TOTP two-factor.

Ships one image, staging first

One container image is built and tested per commit and deployed to staging first. Production only runs the image staging is running, promoted manually, never by a push. The server accepts connections only through Cloudflare, and secrets and dependencies are scanned in CI.

Built with mainstream tools

Next.js, React and TypeScript on a PostgreSQL database, styled with Tailwind CSS and shipped in Docker containers.

Connected sources

Bring your search and ad data next to what AI says.

Your search and ad accounts already show where people find you. Connect them to Vimarsh and read them alongside your AI visibility, so every gap has a source you can check.

Logos are trademarks of their owners and show platforms Vimarsh connects to. Their use does not imply endorsement or partnership. Connect one from your dashboard once it is available to your plan.

Vimarsh Vidyalaya

Making Indians ready to be a Marketing Engineer.

A practical, self-paced learning platform from the Vimarsh team -- real examples and hands-on activities for developing modern marketing, technology, AI, data and automation skills.

Explore Vimarsh Vidyalaya

What you get

What you can do with Vimarsh.

The short version, in the order you'd use it: from describing your business to deleting your data, you stay in control.

  1. 01

    Describe your business

    Name, website, industry, location and services, plus up to 10 competitors.

  2. 02

    Get buyer-style queries

    Generated from your profile across eight kinds of question, with localized variants for Indian locations. Regenerate any time.

  3. 03

    Measured every week

    Vimarsh runs your queries each week on the engines your plan includes.

  4. 04

    See where you stand

    Your latest AI Visibility, mention, recommendation and citation rates, with the sample size behind them.

  5. 05

    Open the response behind any number

    Every AI response is listed by engine and query. Open one to read the raw answer and any sources it cited.

  6. 06

    Choose a plan

    Billed monthly in INR through Razorpay. Your plan sets how many businesses, queries and engines are measured.

  7. 07

    Delete your account

    Confirm with your password. Restore it any time during a 14-day cooldown; after that, your account and business data are erased.

From the blog

Articles on AI visibility, perception intelligence and how machines see your business.

Read the blog

Ready when you are

See what AI says about your brand.

Add your business and the competitors you care about. Vimarsh measures how AI engines describe you, with the response behind every number.

Get your free AI reportSee pricing

Asked through each engine's official API · every number links to its raw answer

नमस्ते · Namaste

Thank you, in the languages of India

  • धन्यवादHindi · Marathi · Nepali
  • ধন্যবাদBengali · Assamese
  • ధన్యవాదాలుTelugu
  • நன்றிTamil
  • આભારGujarati
  • ಧನ್ಯವಾದಗಳುKannada
  • നന്ദിMalayalam
  • ਧੰਨਵਾਦPunjabi
  • ଧନ୍ୟବାଦOdia
  • شکریہUrdu
  • Thank youEnglish

With gratitude, from all of us at Vimarsh.