Profound vs Depra AI, and the fact that only one of them will tell you a price

Read on 31 October 2026, Profound publishes no figure at all for a brand buyer and Depra AI publishes a four plan ladder in rupees plus 18 per cent GST, so the real comparison is a sales call against a currency line on your invoice.

AI Knows Us · Updated 2 November 2026

Pick Depra AI if you want to know the price before you speak to anybody, and Profound if your evaluation can survive a demo booking. That is the first and largest difference between these two, and it is not a detail of presentation. Read on 31 October 2026, the brand side of tryprofound.com/pricing carries two plans, Trial at Free and Enterprise at Custom, and no dollar, pound or euro figure anywhere in the visible page. On the same day depra.ai/pricing carries four plans with a number against each one, Starter at Rs 1,999 a month, Growth at Rs 7,499, Scale at Rs 12,499 and Enterprise at Rs 25,000, each stated as being plus 18 per cent GST, and the page prints the GST inclusive figure beside it.

What each vendor publishes before anybody picks up a phone

Both pages were fetched on 31 October 2026 and both returned HTTP 200. Everything in this table is from the visible text of those two pages on that date.

Published factProfoundDepra AI
Plans with a number on themNone. Trial is Free, Enterprise is CustomFour. Rs 1,999, Rs 7,499, Rs 12,499, Rs 25,000 a month
CurrencyNo currency appearsIndian rupees, with 18 per cent GST stated separately
The unit being soldPrompts tracked, and credits for its agent workPrompt slots and successful checks, both defined on the page
Entry allowance50 unique prompts, run daily for 7 days16 prompt slots and 2,000 successful checks a month
Engines on the cheapest route inChatGPT, Gemini, Google AI OverviewsChatGPT, Gemini, Perplexity, Google AI Overviews, all daily
Engines at the top tierUp to 9 Answer Engines trackedFour, with every other model as a paid Enterprise add on
Exports on the entry routeListed as None, alongside History None and API NoneCSV download, and an MCP server from the first plan
SeatsUnlimited on both plansNot published as a seat limit
Single sign onNo on Trial, Yes at Enterprise with SOC 2 namedNot published
Self hostingNot publishedOption to run on your own server, Enterprise, set up by their team

Two things called seven days that are not the same thing

Profound's free route is a window. Its own pricing page describes the Trial as 50 prompts run daily for 7 days, and the comparison grid beside it sets Prompt Frequency on that plan to Daily for 7 days. After the window the tracking is finished unless an Enterprise package is agreed. Depra's seven days is a card free opening on a plan that then continues: the page states that Starter, Growth and Scale start with 7 days free, no card needed, and that Enterprise trials are opened by the Depra AI team.

That difference decides what each trial can prove. Seven days of daily checks on 50 prompts is a snapshot of a market, and it is a good one, because 50 prompts is a wider set than several paid tiers elsewhere in this category will sell you. What it cannot do is show you movement, because Profound's grid lists History as None and Exports as None on that plan, so the readings do not leave the product and there is nothing to compare next month against. Depra's Starter keeps running at Rs 1,999, and its feature list includes CSV download and the websites AI cites, so the second month is a comparison rather than a repeat.

The engine lists point in opposite directions

Profound's cheap door is narrow and its expensive door is wide. The Trial names ChatGPT, Gemini and Google AI Overviews. Enterprise says Up to 9 Answer Engines tracked, and lists capability for ChatGPT, Perplexity, Google AI Mode, Gemini, Microsoft Copilot, DeepSeek, Anthropic Claude, Google AI Overviews and Exa Search.

Depra is the other way round. Every plan, including the Rs 1,999 one, is published as checking ChatGPT, Gemini, Perplexity and Google AI Overviews daily. Nothing wider is available at any price on the self serve page: the Enterprise panel describes every other model, naming Microsoft Copilot, Claude and Google AI Mode, as an add on arranged by contacting them. So a buyer who needs Copilot or Claude measured gets to that surface through a conversation with either vendor, and the conversation starts from a very different place depending on which one you call.

If you are buying from the United States

Three published details matter and all three are on Depra's own page. The ladder is billed monthly in Indian rupees. The 18 per cent GST line is Indian tax, stated as added on top. Support is email during IST working hours on every plan. None of that makes the product unsuitable, and the page is unusually clear about all of it, including a plain definition of its two units: one prompt slot is one buyer question in one language and one market, and one check is one question asked to one AI engine. It does mean an American finance team is approving a rupee invoice, and a rupee figure read today is a dollar figure that moves with the exchange rate, which is why this page quotes the rupees and does not convert them itself. Depra does publish its own conversion, which is the right way to do it and worth quoting exactly. Its FAQ answers the question how much Depra AI costs in US dollars by naming the rate and the date: at the RBI reference rate on 15 September 2026 of Rs 95.93 to the dollar, Starter is about $21 a month, Growth about $78, Scale about $130 and Enterprise about $261, all before 18 per cent GST, with Enterprise at about $308 including GST. It adds that the amount your bank charges depends on its own conversion rate. That is a dated snapshot attributed to a named source, so it can be repeated with the date attached, and it is not the same thing as this page inventing a rate of its own.

Profound's position is the mirror image. There is no currency problem because there is no published price. What a US buyer gets instead is an Enterprise quote, and the page tells you what that quote is expected to include: SSO and SAML with SOC 2 compliance, all time history, CSV and JSON exports, API access, custom regions and languages, and dedicated support with a named service level option. Its FAQ also answers the setup question in one word, Instant, and states that the self serve Agency Growth plan comes with 400 credits per month per client workspace, which is the only numeric allowance published anywhere on that page.

The rule that decides it

Write down which of these three you are buying, and the answer follows.

  • A documented monthly cost you can approve this week. Depra, and the Rs 1,999 tier is a real plan rather than a teaser, with 16 slots and 2,000 checks.
  • Breadth of engines and a security review your IT team signs. Profound, because SSO, SOC 2, API and all time history are all published as part of the same tier, and Depra publishes none of those four.
  • Agency work with several client brands. Compare Depra's Scale tier, 120 prompt slots pooled across 5 client projects with your logo and colours on the reports, against Profound's published 400 credits per month per client workspace. These two are the clearest published agency positions of the pair, and they are counting different things, so price one quote against one month of your real client list rather than against each other.

Where we sit on the same two questions

We sell in this category, so the same test applies to us. Read on 31 October 2026 our own pricing page publishes a price for every plan, and it publishes one in rupees and one in dollars, which is one more currency than either vendor above. GEO Growth is $299 a month for 100 buyer questions, 1,200 scheduled answers and all five engines including Claude. It also publishes the two places we lose this comparison outright: the page states Logins included, 1, full access, where Profound publishes unlimited seats on both of its plans, and there is no single sign on anywhere on our page, where Profound names SSO and SOC 2 in print. If those two things are on your requirements list, Profound answers them today and we do not. Our cadence is weekly on priority questions, which is slower than the daily checks both vendors above publish, and our own record is six citations in 1,260 readings of the second audit, none of them in the United States.

If your shortlist is genuinely Profound and Depra rather than a wider field, the most useful next read is our page on Depra AI alternatives on a budget or for a small team, because the budget end is where these two stop being comparable at all.

Questions buyers ask about Profound and Depra AI

Does Profound publish any price at all

Not for a brand buyer, as read on 31 October 2026. The brand side of its pricing page carries Trial at Free and Enterprise at Custom, with no currency figure in the visible text. Its FAQ does publish one allowance, 400 credits per month per client workspace on a self serve Agency Growth plan.

Can a company outside India buy Depra AI

The page sells into 40 markets and names India, the US, the UK, Europe, the Gulf and Australia, so the measurement is not India only. The billing is: the ladder is stated as billed monthly in Indian rupees with 18 per cent GST added, and Enterprise is invoiced by NEFT, RTGS or UPI.

Which one tracks more AI engines

Profound at its top tier, which says up to 9. On the cheapest route in it is the other way round, because Depra publishes four engines checked daily on its Rs 1,999 plan and Profound's trial publishes three.

What is a successful check and what is a prompt slot

Both are defined on Depra's own page. A prompt slot is one buyer question in one language and one market. A check is one question asked to one AI engine. So 16 slots on four engines daily is where the 2,000 monthly checks on Starter comes from, and that arithmetic is published rather than inferred.

Is a free trial enough to choose between them

It is enough to answer different questions from each. Profound's week of daily runs on 50 prompts tells you what the engines currently say about your market. Depra's seven free days on a plan that continues tells you what the reporting looks like in the second month, which is the month that matters. Run both and keep the exported files, because only one of the two trials produces any.

Does either one commit to a security standard in writing

Profound does, naming SSO, SAML and SOC 2 compliance on its Enterprise tier. Depra publishes a different kind of assurance, an option to run the product on your own server, set up by its team, on Enterprise. Those answer two different procurement questions and a security review will ask for both.

How we know this

We read the assistants rather than guessing what they do. In our own audit of 1,260 saved answers from ChatGPT and Gemini across the United States, the United Kingdom and India, 82.0% of every citation landed on a site the business itself owned, and review directories took 0.7%. Counted as one domain per reading, 16,516 of 20,143.

That is the whole asymmetry: the pages you own are the pages an assistant quotes back to a buyer. See the dated readings and the method on our proof page.

Find out whether assistants name your business, on your own domain and your own buying questions.

By Aditya Goel, Founder, AI Knows Us. Reviewed 1 November 2026.

Read next

AI Visibility Tracking Tools in the UK: Twelve Vendor Pages Fetched, Five Surfaces Named in the Markup, and Two With No Price In ItOn 16 October 2026 we requested twelve pages belonging to six vendors in this category and ran no JavaScript on any of them. Every money figure we could extract was in dollars, there was not one pound figure in the set, and two of the pricing pages carried no money figure at all.Profound pricing, the trial limits it publishes and the agency plan named only in its FAQ, compared with six priced rivalsThere is no currency figure anywhere in the document the server sends, and the two numbers it does publish are a seven day trial and 400 credits a month per client workspace.Profound alternatives on a budget or for a small team: its published ladder is a seven day trial and a quote, and when a buyer says money the engines answer with a free Google productRead on 21 October 2026, tryprofound.com/pricing publishes two plans and neither one is a monthly subscription a small team can buy, and in 28 of our budget framed readings only 8 put a paid tool in this category first.Carpet cleaning advertising, the four channels that produce calls and the Google product that bills you per leadCarpet cleaning services is a named Local Services Ads category, read from Google's own help page on 31 October 2026, which changes the order a cleaning business should spend in.Marketing certifications compared on what each issuer publishes before you pay, with the prices read on 31 October 2026Four issuers, one criterion that anybody can check, and the two that publish no exam price on their own pages at all.Ecommerce consulting firms, the four engagements they sell and how advisory differs from hiring an agencyWhat a consultancy is for when the problem is the decision rather than the execution, the four shapes the work takes, how each is priced, and the deliverable to insist on.

Or browse the whole archive.

See what AI says about you.

The first scan is free, it needs no account, and it asks the questions your buyers ask.

Type your website address to continue.

Free. No card. We ask 5 real buyer questions on 2 AI apps.

Already sure? Start free. No card needed, and the free plan stays free. Pay only when you want the work. Or talk to a person.

Or read how we fix what it finds.