Harsh Vitra - AI Product Founder

Founder @brandlabs.app ★

Co-founder @enrootmumbai ★

Designer | Innovator

[Harsh Vitra - AI Product Founder @HarshVitra]

@Divyyanshishrma Sorry but I’m seeing this for the 10th time now reposted repeatedly

Getting irritated

[Harsh Vitra - AI Product Founder @HarshVitra]

3/3
Because of these pipelines they’re loosing a high intent high urgency buyer
Zerodha made stocks easy, even though the govt had several steps and processes
Someone needs to do this for vehicles

@anandmahindra what do you think?

[Harsh Vitra - AI Product Founder @HarshVitra]

1/3
Sometimes complicated systems make business much more difficult than it should be

Eg. I have decided to buy XEV 9s and ready to make the payment, all I wanted to know was delivery timelines of different dealers to book the fastest one

Reached out to 3 dealers and…

[Harsh Vitra - AI Product Founder @HarshVitra]

2/3
Reached out to 3 dealers and none of them replied straightforward, now my query is sitting in 3 different stages
Dealer 1: wants me to take a test drive - I don’t need it
D2: the sales guy will be reaching out to me, they didn’t
D3: have uploaded my enquiry, 0 updates

[Harsh Vitra - AI Product Founder @HarshVitra]

I ran Brandlabs for 22 months in San Fran Sim — peak $881k MRR, then we IPO’d. Same map, seed 2515190445: https://t.co/av9ewjRNtk

Crazy game

[Harsh Vitra - AI Product Founder @HarshVitra]

Today’s AI generated vibe coded products are the worst set of products coming out in the market.

Flashy launches, hyped marketing and dead in a few months

Don’t make important decisions based on the viral content you see online by influencers who haven’t built anything real

[Harsh Vitra - AI Product Founder @HarshVitra]

I’m Curious

What tools or software are you using for your startup?

Always looking to discover better workflows.

And if you’re just starting to build, let’s connect.

Happy to share what’s worked (and what hasn’t) so you can move faster and build for scale.

[Harsh Vitra - AI Product Founder @HarshVitra]

Being opinionated about your tech stack will save you a lot of time.

Also, having a clear workflow makes building startups much easier.

Here’s my end-to-end workflow for taking an idea to production 👇

[Harsh Vitra - AI Product Founder @HarshVitra]

There are obviously many more tools.

The stack changes depending on the product you’re building.

This is simply the workflow that has worked consistently for me.

No brains, 0 to 1 quickly

[Harsh Vitra - AI Product Founder @HarshVitra]

Latest addition: @TermwiseAi

It scans my Cursor repository and tells me whether the code is actually production compliant.

With vibe coding becoming common, tools like this are becoming increasingly important.

It’s still in beta.

Let me know if you’d like to try it.

[Harsh Vitra - AI Product Founder @HarshVitra]

@TallyForms

Whenever I need forms.

Fast to build, easy to connect with workflows, and works really well with n8n.

[Harsh Vitra - AI Product Founder @HarshVitra]

@n8n_io

Whenever I need workflows or quick automations.

Apparently I’ve been using it since 2019, long before AI made it mainstream 😄

[Harsh Vitra - AI Product Founder @HarshVitra]

@metabase

For visualising and analysing transactional data directly from the database.

Simple, fast, and gets the job done.

[Harsh Vitra - AI Product Founder @HarshVitra]

Now for the product stack 👇

@posthog

Everything related to product analytics.

I still use GA where Google Ads requires it, but everything else runs through PostHog.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 6: Deployments

Frontend → @vercel

Backend → Render

Once the product scales, I move the backend to @awscloud.

Simple.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 5: PostgreSQL (via @supabase)

I use Supabase as a managed Postgres db.
0 SDK in my code

Direct db connection so I can migrate, no vendor lock-in
Same idea for storage, I use the S3 buckets
Supabase is love ❤️

But at scale & for compliance, we usually end up moving later.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 4: @nestframework

For the backend, I use NestJS with TypeScript + Sequelize.

Clean architecture, scalable codebase, and easy to maintain as the team grows.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 3: @nextjs

The first task I give Cursor is migrating the frontend to Next.js.

Vite is great to start with, but it won’t take you very far for a production product.

You’ll thank me later.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 2: Cursor

I connect Git to Lovable, clone the repo, and disconnect Lovable.

The first thing I do inside Cursor?

Delete anything that says gptengineer or lovable.

Clean slate.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 1: Lovable

I use Lovable to build the first prototype.

It’s fast, the visuals are good, and it helps validate ideas without wasting engineering effort.

[Harsh Vitra - AI Product Founder @HarshVitra]

Step 0: @ChatGPTapp

I start by refining my PRD.

The better the PRD, the less confusion later. Spending time here saves days of rebuilding.