CreatOK is an AI TikTok e-commerce video generator

You open this post with the models you are built on (Soar, VEO, Kling, and so on) and "turn one winning video into hundreds," and both undersell you. The model list invites the obvious reaction, "I could just use those directly," and volume is a commodity: every AI video tool makes bulk clips, and a TikTok Shop seller who floods the feed with near-identical videos can tank their own reach.

Your real edge is four lines down, in the Video-to-Video workflow: discover a viral hit, deconstruct it, fuse the key elements, produce a new asset. That is a system for reverse-engineering proven winners, and no raw model hands you that. Lead with "find what is already winning on TikTok and rebuild it for your product," put the GMV number from your site right next to it, and let the model list and "hundreds of videos" be the how.

Which feature do your 300K sellers actually come back for, the volume, or the viral discovery?

You Should try StartupSubmit and Thanks me later

Hey ,

I'm Fenil — a full-stack developer from India. I build ERP and SaaS systems professionally, and I've been sitting on a problem that I personally face every single day.

The problem

India processes 20 billion UPI transactions every month. GPay, PhonePe, Paytm — everyone uses them. Every transaction records the amount and who you paid. None of them record why.

A month later, ₹180 to some UPI ID means absolutely nothing. Was it a team lunch? A petrol stop? A friend you lent money to? That context is gone the moment you close the payment app.

I personally solve this by opening my Notes app 2–3 days later, cross-referencing my bank statement, and trying to reconstruct why I paid. It takes 20 minutes and I still get it wrong sometimes.

I know I'm not alone because every person I've shown this problem to has said "yes, I do the exact same thing."

What exists today — and why it falls short

There are apps that read your bank SMS and auto-categorize transactions — Walnut, MoneyView, ET Money. I've tried them all.

The problems:

• They track what you spent. None of them capture why.

• Most have quietly pivoted to selling you loans using your financial data. Expense tracking became the bait.

• Users don't trust them anymore — and for good reason. Your SMS inbox contains OTPs, salary alerts, account numbers. Handing that to an app that monetizes your data is genuinely risky.

• The UIs feel like they were built in 2015 and never touched again.

The insight I keep coming back to

The ledger problem is solved. The story problem isn't.

₹180 to Zomato is a data point. ₹180 to Zomato on a Tuesday evening because your team stayed late and you ordered for everyone — that's a story. That's something you'd actually remember, learn from, and care about at month end.

Nobody has built the context layer on top of Indian UPI transactions.

What I'm thinking about building

An app that:

• Auto-reads bank SMS the moment a transaction arrives (works across all UPI apps — GPay, PhonePe, Paytm, everything — because the bank SMS is universal)

• Fires a notification within 3 seconds: "₹180 paid — what was this for?" with one-tap answers

• Builds a personal spending story over time — not just a ledger

• Does all SMS parsing on-device. The raw SMS never leaves your phone. Only structured data (amount, merchant, date, your tag) syncs to server — encrypted.

• Makes money through subscriptions, not by selling your data — structurally aligned incentives

Think Spotify Wrapped but for your spending. Every month, a beautiful personal story of where your money went and what it meant.

What I'm genuinely uncertain about — and why I'm posting

I'm not here to pitch. I'm here because IH has people who've actually shipped consumer apps and can stress-test thinking better than anyone I know locally.

Three things I'd love honest perspective on:

1. Is the 'why' problem real outside India? I'm assuming this is universal — anyone using digital payments loses context over time. But maybe Indians are uniquely bad at this because UPI adoption happened so fast that habits haven't formed. Would love to know if you've felt this in your market.

2. What actually makes a daily financial habit stick? I've thought about post-payment prompts, evening digest notifications, and a monthly Rewind card. But I know habit formation is brutal in consumer apps. What's the one mechanic you've seen actually work?

3. The SMS permission trust problem Asking users for SMS READ permission in India right now is genuinely hard — previous apps damaged that trust badly. My plan is on-device parsing + open-source the parser + zero OTP capture by design. Has anyone navigated a similar trust deficit with a new app in a category that bad actors already tainted?

Would love honest reactions — including "this is a bad idea because X." That's more useful to me than encouragement right now.

Thanks for reading this far.

— Fenil

P.S. If you've built anything in Indian fintech or consumer habit apps, I'd especially love your take.

Turning one winning TikTok into hundreds of variations is a compelling idea. The no-prompt workflow makes it accessible to non-creators too. Would love to see some real-world case studies showing ad performance improvements.

The stronger angle here is not really “anti-AI homogenization.” It is proprietary signal infrastructure for AI content and commerce teams.

If Keyapi can unify TikTok, Amazon, Instagram, YouTube, Shopify, CRM, and campaign data into one clean layer, that is much bigger than a content tool. It becomes the private context layer that makes AI outputs sharper because the model is grounded in data competitors cannot see.

The part I would pressure-test is the name. Keyapi.ai is clear for an API product, but it also boxes the product into a technical key/API frame. If the bigger category is private commerce signal, brand context, and AI-ready data infrastructure, the product may need a name that carries more category weight before more integrations, docs, and customer memory lock in.

Exirra .com would fit that direction well because it feels broader than an API utility and stronger for AI signal intelligence, discovery data, content context, and private-data infrastructure.