Video summary
How This 26 Yr Old Makes ₹20 Lakhs Every Month By Building AI Apps | #246 The Sanskar Show
Main summary
Key takeaways
Core business claim / outcomes
- Built multiple AI-powered apps with very low/no traditional coding effort.
- Emphasizes repeatability via:
- Market demand
- Rapid shipping
- Revenue reported:
- ₹15–20 lakhs/month
- One app reportedly made ₹20 lakhs every month
- Scale reported:
- Built 15 applications by “typing English into a text box” (no code / minimal code).
- Bootstrap model:
- Raised zero funds so far
Product strategy: “don’t educate the user; build for existing demand”
Core preference
Build products where:
- The problem already exists
- The market already understands the need
- Users are already paying for a workaround/alternative
Avoid
- Requiring user behavior change
- Products that need long “education” cycles (slower path to profitability; not bootstrap-friendly)
Heuristics / positioning playbook
- Build a better solution than what exists, not a totally new category.
- Use “paid-only” early:
- Require customers to pay to confirm willingness-to-buy.
Idea selection & market research loop (MVP + paid validation)
Market research approach
- Identify what’s working on existing platforms (e.g., Twitter, app marketplaces).
- “Put it out there” and watch reactions.
MVP validation loop
- Build an MVP quickly
- Test market response
- If users show excitement and/or are ready to pay, continue investing time
Direct thesis: willingness-to-pay
- Use willingness-to-pay as the strongest signal:
- “Thousands of people say ‘good’” is weaker than people who actually pay money.
Customer acquisition & content distribution system (multi-channel, multi-product)
Content creation strategy
- Posts continuously (“post wherever you can for free”):
- YouTube
- Facebook groups
- etc.
SEO / platform targeting strategy
- Creates separate YouTube channels per product to avoid confusing the audience/algorithm.
Content volume tactic
- Example reported behavior:
- Posting 17 Instagram videos in a day
- “Worst case some don’t work—next day post more.”
Marketing objective
- Reduce dependence on “perfect hooks/end” by:
- Increasing content throughput
- Iterating based on performance
Tool-assisted competitive analysis (App intelligence)
App Arc platform usage (as a sourcing/validation tool)
- Purpose:
- Find “top performing apps”
- Infer monetization + competitor landscape
- Example filters/inputs:
- Region: India
- Category/tags (e.g., Art & Design, Sports → Cricket)
- Time window: June 2025 to July 2026
- Revenue range: $100 to $2,000
- Output includes for target apps:
- Revenue last 30 days (example shown: $2,000)
- Downloads
- Promotional pages, ratings/reviews, description
- Version history (what changed and when)
- In-app purchase ranges
- Competitor apps + update monitoring
Business use
- Reverse-engineer product-market fit by:
- Estimating monetization potential
- Identifying what customers value
- Using competitor update history as a “product roadmap signal”
Execution workflow: build → deploy → ship fast (AI-driven dev ops)
AI development approach
- Uses a Claude + terminal workflow.
- Connects AI to tools like a Chrome extension to directly operate browser actions.
- Principle:
- Trust AI for the “how”
- Focus on the end result
Local-to-live deployment
- Hosting preference mentioned:
- Cloudflare (compared to Vercel as “cheap”)
- Deployment tooling:
- Wrangler CLI to authenticate and deploy
- Deploy locally running app to Cloudflare and share the resulting URL
Concrete “build from scratch” case: online multiplayer game (Grid Rush / Snake & Ladders variant)
Process steps described
- Use terminal to scaffold a new project/folder (with AI).
- Ask for a multiplayer web game and target hosting.
- Later hosted on Cloudflare
- Iterate with AI prompts:
- Request design changes (e.g., replace gradient/logo style, improve dice look, move cars, color changes)
- Deploy to Cloudflare and test via shared room links
Multiplayer / game operations concepts
- Room-based joining
- Synchronized turns
- Running local dev + live deploy URL testing
- “Parallel work” mindset:
- Multiple terminals/tasks running (e.g., 10 terminals) until a blocker occurs
Iteration philosophy
- Get the product live first; improve later once traction/feedback exists.
- Acknowledges UI/design shortcomings but treats design as an iteration backlog (not a launch blocker).
Product catalog / examples of app types (what he ships)
Examples (earlier and later)
- Magic Slides (main product)
- Presentation creation
- Built related “editor” capabilities
- Later reimagined variants
- Blur Web / Blur Screen
- Browser extensions to blur sensitive content
- MR Track
- Revenue tracking app
- Built an internal tracking version after Razorpay connection issues
- PPD GPT / GPD
- Content/publishing automation around presentations (described as “take on my own product”)
- Build Cheap (pricing strategy shift)
- AI tools where users pay directly per action (avoid confusing “credits”)
- Uses OpenRouter models and charges very low amounts (examples: ₹2/₹5/₹10 style)
- Offline/local AI
- Offline LLM app supporting multiple models (chat + video understanding mentioned)
- Magic Chat
- “Knowledge bot” that takes PDFs/URLs/YouTube into one Q&A system
- Video creation workflows
- Tools described like:
- Hyperframe (HTML → video)
- Chatterbox (open-source/local audio generation)
- Supports vertical/horizontal content, scripts, voice, edits, captions, etc.
- Tools described like:
Monetization & pricing plays (tactics to reduce conversion friction)
Paid-only thesis
- Don’t develop while users are only “interested.”
- Require payment to validate demand.
Credits vs transparent pricing
- Criticizes credit-based pricing (“fake currency”).
- Recommends transparent pricing:
- “See the money directly” per request.
Lifetime deal as early differentiation
- Example edge:
- Launched early with a $7 lifetime deal instead of recurring subscriptions.
- Rationale:
- Lifetime deals convert easier for beginners
- Reduces purchase-risk and “uncertainty” barriers
Metrics & KPIs mentioned (explicit and implied)
Explicit metrics/claims
- Revenue target/claim:
- ₹15–20 lakhs/month
- “₹20 lakhs every month” for one app
- App-building scale:
- 15 apps live
- App Arc examples:
- Example: $2,000 revenue in last 30 days
- Another example: $980 revenue and 995 downloads in last 30 days
- Marketing throughput KPI:
- 17 Instagram videos posted in a day
Implied/operational KPIs
- Downloads → retention
- Mentions a trend that more apps were built, but retention was declining.
- Paid conversion
- Willingness-to-pay is the gate.
- Content iteration loop:
- Post more when performance is weak.
Strategic framework references (implicit) + playbooks extracted
- Paid MVP validation
- Build for existing demand
- Launch 10, expect 9 fail (portfolio mindset)
- Content throughput as iteration leverage
- Separation of audiences by channel
- One product → one channel strategy
Actionable recommendations distilled from the talk
- Validate ideas with money, not compliments:
- Charge early; ensure customers pay to confirm real demand.
- Choose problems where users already know the solution:
- Beat existing alternatives rather than educating the market.
- Use competitor/app analytics:
- Track revenue, downloads, reviews, and version changes to identify what’s working.
- Build and deploy quickly:
- “Get it live” first; treat UI as an iterative backlog.
- Market continuously and broadly:
- Publish to multiple channels; create separate brand spaces per product.
- Use transparent pricing or low-friction deals:
- Prefer action-based pricing over credit confusion.
- Consider lifetime deals early for faster conversion.
- Adopt a portfolio execution model:
- Build multiple simple products to discover the one that “hits.”
Presenters / sources
- Presenters: Sanskar (guest) and the host (interviewer; name not stated in subtitles)
- Source/tools mentioned:
- App Arc
- Claude
- Cloudflare
- Wrangler CLI
- OpenRouter
- Hyperframe (HTML → video)
- Chatterbox (audio generation)
- Additional references:
- Twitter, Reddit, Product Hunt, App Store/Play Store, Indie Hackers community