He squats Google's marketplace and prints $17K/mo

The Case: $10K MRR steady, $17K peak month, 0 paid acquisition

What he does. Sells an AI tool that turns a topic, a YouTube link, a PDF, or raw text into a styled PPTX or Google Slides deck.

Backstory. Self-taught Indian dev, started in a call center, has shipped 24+ small products under "IndianAppGuy." Around the GPT-3.5 launch, someone DM'd him asking him to build a pitch deck. He turned the gig down and built a tool that would. Shipped v1 in days, listed it on the Google Workspace Marketplace, then let Google's own search inside Slides do the customer acquisition.

Product. - Web app + Google Slides add-on + ChatGPT plugin + Figma integration - Inputs: text, topic, URL, YouTube video, PDF, DOCX - Outputs: PPTX, PDF, Google Slides - "Clone Slide" — image to editable PPT (Premium tier) - Self-serve, no sales calls, no onboarding

Pricing. Free (3 decks/mo, watermark) → $8/mo Essential → $16/mo Pro → $29/mo Premium. Annual ~22% off. Separate dev API tier.

Key numbers. - Steady-state MRR (own IH report): $10,000/mo - Peak 30-day revenue (Aug 2024): $17,000 - Lifetime presentations generated: 10M+ - Users across web + add-on: 2M+ - Google Slides add-on installs alone: 1M+ - Paid acquisition spend: $0 - Portfolio bootstrapped ARR (8 products): $300K - Lifetime portfolio revenue (verified): ~$809K - Entry-paid plan price: $8/mo (coffee, not SaaS) - Time from DM to v1: days

Stack. OpenAI API, Google Workspace Marketplace add-on, Next.js web app, Stripe, blog on subdomain, Loom for Reddit videos, YouTube channel as long-tail SEO funnel.

Why this works (and what most readers will miss)

1. The marketplace listing is rented distribution that ranks inside Google's own UI. When a teacher inside Google Slides types "AI" into the add-on search, MagicSlides surfaces. That's not SEO, it's not ads, it's a slot inside the workflow itself. Reviews compound. New entrants start at zero. This is the entire moat, and it costs $0/mo to maintain.

2. The $8 entry price is the conversion engine, not the AI. Most "AI deck" tools price at $19-29/mo because some growth blog said so. He priced at $8 because that's a Spotify, not a SaaS commitment. Free → $8 is the only friction point in the funnel and he sanded it down to nothing.

3. He cut every channel without attribution. Twitter has 24K followers and didn't move revenue. Facebook groups didn't either. Reddit how-to videos and YouTube tutorials did, because installs and Stripe checkouts trace back to them cleanly. Most operators run all five channels because gurus tell them to. He runs the two that show up in his dashboard and ignores the rest.

The Distilled Read

The product itself is a weekend. OpenAI does the slide generation. He admits this openly. Anyone reading this can ship a clone in 72 hours: text input → GPT prompt → python-pptx → file out. The hard part of MagicSlides isn't the code, it isn't even the prompt engineering. The hard part is being on the Marketplace listing two years before you. That's not a software problem, it's a queue problem. If your instinct on this case is "AI wrapper, easy" — you've already missed it. The wrapper is supposed to be easy. The wrapper is the giveaway, not the business.

One playbook, run 24 times. Notice he didn't bet his life on MagicSlides. He shipped it as one of two-dozen small tools and let the market vote. SheetAI, BlurWeb, BlurScreen, MagicForm, MagicChat — same template, same architecture, same Stripe integration, same Reddit launch motion. Twenty-three flopped. One prints $10K/mo. That's the only honest way to play AI tools right now: clone your own template, ship the next one before the last one is finished, let the marketplace tell you which idea was correct. If you're agonizing for three weeks over which idea to pick, you've already chosen wrong — pick three, ship three, kill two.

The economics only make sense as a portfolio bet. Look at the numbers from the side: $300K ARR across the portfolio, $10K/mo from the breadwinner, the rest scrap. If you treat MagicSlides as a 1-of-1, the survivorship bias eats you. If you treat it as 1-of-20, it's a healthy outlier with capped downside (a weekend's worth of code) and uncapped upside (a marketplace slot that compounds for years). Capital at risk: a domain and your time. Time-to-revenue: weeks. Decay risk: the one to watch. Google could ship native AI in Slides tomorrow and the marketplace traffic halves overnight. That's the bet. He took it knowing the half-life.

Distribution is downstream of attribution, not effort. The lesson buried in his channel choices is colder than it looks. He didn't pick Reddit and YouTube because they were hot, he picked them because both have a pixel-trail back to a paid signup. Twitter has a vibe. Reddit has a referrer. Pick the channel where you can prove the ROI in your own dashboard, not the one where the influencer told you "just show up consistently for 18 months." Cold-DM closes one in thirty only when you can count to thirty. You can't count to thirty on Twitter.

Here is what nobody in the case noticed. The Google Workspace Marketplace listing is not a moat. It is a lease. Google is shipping Gemini-native AI inside Slides as we speak — the day a "✨ Generate slides" button lands in the toolbar, marketplace traffic to MagicSlides decays in a month, not a year. The actual moat being built right now is the 2M installed users, and only if he is capturing them. The operator who survives 18 months from today is the one who, this quarter, started capturing emails on signup and shipping a second product to that list — a Notion-deck tool, a Loom-summarizer, a Figma-to-deck converter, anything that lets him say "you used MagicSlides, here's the next thing" before Google eats the listing. The marketplace is the customer acquisition channel. The email list is the business. Sanskar's portfolio instinct hints he sees this. Most operators copying him won't. They'll celebrate the listing the day they get in and grieve it the day Google ships native.

Steal-the-Playbook

1. Pick one workflow inside one platform, not the category. Not "AI tools." Pick "AI inside Notion" or "AI inside Shopify" or "AI inside Webflow." The marketplace slot is the prize, the AI is the excuse to claim it.

2. Ship a v1 in 7 days. Next.js + OpenAI API + Stripe + the platform's export format. No team plans, no SOC2, no auth flexing. Submit to the marketplace on day 8 and start the review queue clock — that queue is your competitive moat against the person reading this newsletter tomorrow.

3. Price entry at $7-9/mo, not $19. Free tier with watermark, paid tier at coffee price. Free → $8 is the entire funnel. Pro and Premium are upsells for the 5% who hit the cap. The $8 plan pays the bills.

4. Run two channels with attribution, ignore the rest. Reddit + YouTube, or Pinterest + SEO blog, or marketplace + niche newsletter. Rule: if you can't see the install land in your dashboard, it didn't happen. One Loom per week per subreddit, one tutorial per week on YouTube, for six months. That is the work.

5. Capture the email on signup from day one. Welcome email, weekly tip, soft pitch for product #2 by month four. The marketplace giveth users and Google taketh them — the inbox is the only place you keep them.

Stack: Next.js, OpenAI API, Stripe, Loops or ConvertKit, Google Workspace add-on SDK. ~$50/mo to run. ~10 days to first paying user.