AI study tool (formerly TurboLearn): upload class recordings, PDFs, YouTube videos or notes to generate notes, flashcards, quizzes and podcasts.
Two freshmen who started in 2023 (Sarthak Dhawan at Duke and Rudy Arora at Northwestern, childhood friends from a Dallas suburb) built "record the lecture, get notes/flashcards/quizzes automatically" as TurboLearn. The first 100 users came from handing out donuts in the Duke freshman dining hall and posting flyers in bathroom stalls. The real takeoff came from Arora himself filming "how I use this app" TikToks in classrooms, several a day, most of which nobody watched until one blew up (self-reported about 100 million views). They then copied that format across 101 owned accounts and 2,300+ videos, and used the TurboShip program to pay outside creators to post from fresh accounts. Before running any ads they already had 300K users and $500K in revenue. In 2025 it renamed to Turbo AI and moved to turbo.ai, and in 2026-07 it passed 10 million users.One-line growth story, compiled by us from public sources; speculation is marked separately below.
The main engine is a short-video matrix: 101 owned accounts plus paid creators post videos, direct opens are 51%, search is all branded, and people search the name after seeing a video.
The turbo boost was the founder's own viral TikTok, with the first 300,000 users and $500,000 in revenue both coming before paid ads; the lubricant is the daily need to review, with returning users coming back to record lectures and flip flashcards.
By estimate, over 40% of visits are returning users coming back to review; about 5 clicks into the checkout page per 100 visits.
Model: Lenny's racing growth framework. Engine = a short-video matrix (owned accounts and paid creators keep posting "how I use this app" videos, and viewers search the name and visit the site); turbo boost = the founder's viral TikTok and the later controversial videos; lubricant = daily use for classes and exams, with returning users coming back again and again. It got started with the first 100 people brought in through on-campus outreach at Duke.
Real data: 2,000,000 visits, month-over-month +27.24%, largest channel direct at 50.98%, top three channel ranking (direct, organic search, referrals), organic search 100% within search, top search terms (turbo ai, turbolearn, turbolearn ai, turboai, turbo learn), referring sites (simplycodes.com 1.12%, google.com 0.73%), university sites at 15.8% in referral categories, social sources (YouTube, Instagram, X), countries (US 37.93%, India 10.35%, Philippines 8.56%, Canada 6.63%, UK 3.29%), bounce rate 36.15%, 4.56 pages per visit and average time on site 5 minutes 39 seconds come from the public Similarweb page for last month; the public page does not give the shares of the second and third channels. 98,854 visits (about 98,900) to the Stripe checkout page come from quietmoney checkout-page data.
Supporting evidence: Made by two freshmen from the class of 2023, originally named TurboLearn and renamed Turbo AI in 2025; the first 100 people came from handing out donuts in the Duke dining hall and flyers in bathrooms; Arora filmed a TikTok in a classroom and one went viral (founder says 20M to 100M views), then it was copied to 101 owned accounts and 2,300+ videos, and the TurboShip program pays outside creators; before running ads it already had 300,000 users and $500,000 in revenue; in 2025 it had 5 million users and eight-figure ARR, and in 2026 users passed 10 million with cumulative revenue of about $13 million.
Inference: The two channel circles are real; how each route is split is an inference: direct 1,019,600 = returning users 600,000 + classmate recommendations 219,600 + seeing short videos 200,000; the other channels 980,400 = searching the name after seeing short videos 450,000 + returning users searching the brand name to log in 250,000 + school sites and coupon code sites 150,000 + classmate recommendations 130,400. Semrush desktop data shows direct 68.5% and search 31.5%, which suggests search is the bigger part of the second circle, so most visits brought by short videos are placed on the search side. The rates of 3% visiting after seeing short videos, 40% opening after being recommended by classmates and 10% clicking through after seeing it on school sites and coupon code sites are guesses.
Checkout: checkout.stripe.com
“Study with friends, use pre-made study guides, and learn on the go (no invite flow shown); 'generous free tier that includes note generation, flashcards, and qu”Quote from the site
New-user source: Other (Direct / referral / email / paid). Direct 68.5%, search 31.5% (brand keywords 100%: turbo ai, turbolearn); referral simplycodes.com 1.12%; social sources YouTube, Instagram, X; AI referral ChatGPT 0.07%. New users all land as direct/brand visits. last month 99K, +109% month over month.
Semrush source breakdown (desktop)
| Attributable new users, last month | Share |
|---|---|
| Other | 100% |
Direct + brand search 100%. Direct and brand search count as viral only when the site shows sharing mechanics; the largest share wins, under 5% means Other.
| Country | Share of visits |
|---|---|
| United States | 37.93% |
| India | 10.35% |
| Philippines | 8.56% |
| Canada | 6.63% |
| United Kingdom | 3.29% |
Countries (Semrush): United States 27.3%, Bangladesh 12.2%, India 8.8%, United Kingdom 5.9%, Canada 5.7%
English
YouTube, Instagram, X (Twitter)
| Source | Share |
|---|---|
| ChatGPT | 0.07 |
| tabbit.ai | 0.03 |
| Gemini | 0.02 |
AI referrals (Semrush): ChatGPT 1,199 (0.07%), tabbit.ai 448 (0.03%), Gemini 379 (0.02%), leonardo.ai 232 (0.01%)
1.9K keywords · brand ~100% · organic 100% / paid <0.01%
| Similarweb top 5 keywords |
|---|
| turbo ai |
| turbolearn |
| turbolearn ai |
| turboai |
| turbo learn |
5 referring sites
| Site | Share of referrals |
|---|---|
| simplycodes.com | 1.12% |
| google.com | 0.73% |
| Category | Share of referrals |
|---|---|
| Computers Electronics and Technology - Other | 24.63% |
| Universities and Colleges | 15.8% |
| Digital Marketing | 1.12% |
| Search Engines | 0.73% |
About 98,854 last month (Similarweb estimate), #15 of 606; the month before was 50,797. Opening checkout is not paying, so this is not revenue.
New-user source: Other (Direct / referral / email / paid). Mostly referrals, email or paid, or mostly direct with no clear source Direct 68.5%, search 31.5% (brand keywords 100%: turbo ai, turbolearn); referral simplycodes.com 1.12%; social sources YouTube, Instagram, X; AI referral ChatGPT 0.07%. New users all land as direct/brand visits. last month 99K, +109% month over month.
Free tier includes note generation, flashcards and quizzes; paid upgrade price Not found
Dossier confidence: medium. Parts 1 and 2 come from reading the site and Similarweb data; part 3 from public reporting, founder posts, Wayback snapshots, WHOIS and similar, with speculation marked.
The number: checkout visits is Similarweb's estimate of visits from the site to a checkout payment page; opening checkout is not paying, so this is not revenue. Small sites land on fixed buckets.
Data through last month. Page updated 2026-10-31.