Prajwal Tomar Profile picture
Helping founders build AI-powered MVPs fast. Building @clarity_board (SaaS), @aimvpbuilders (community) & @ignytlabs (agency)
13 subscribers
Jul 17 10 tweets 2 min read
Save this if you want your next AI site to ACTUALLY look good.

This is the exact UI workflow I use on every client MVP 👇 1/ Stop saying “make this UI beautiful.”

That’s exactly why your site looks messy and inconsistent.

Here’s what I do instead:
→ Take a screenshot of any site I love
→ Drop it into ChatGPT
→ Ask it to create a design.json that captures the layout, colors, spacing, typography

Then I tell Cursor:
“Use this design.json for styling only.”

Works insanely well.
Jul 15 14 tweets 3 min read
Want to know how I use Cursor and Lovable to ship MVPs fast for clients?
I’ve built 18+ products using this exact workflow.

Here’s the full system I use to plan, design, build, and launch.
👇 Save this. You’ll need it. 1. Start with planning

Before writing a single line of code, I plan everything inside ChatGPT.

The documents I create:
- PRD (Product Requirements Document)
- UI Development Plan
- Database Design
- Implementation Plan

This step is where 90% of the clarity comes from. The more structured these are, the better the output across every tool.
Jul 11 15 tweets 3 min read
I stopped using Cursor and switched to Claude Code.

Best decision I’ve made in a while.

Here’s why I’m not going back (and the tips saving me 10+ hours every week) ↓ 1/ Cursor kept messing up my bigger projects

Whenever I worked on large codebases, Cursor started breaking.

Context got confusing. Code suggestions didn’t make sense.
I was fixing AI mistakes more than building.

It got frustrating.
Jul 8 12 tweets 3 min read
I ditched “vibe coding.”

And suddenly, building with Cursor felt calm again.

Here’s how I ship faster without burning out ↓ Reposting this thread because many of you didn’t catch it the first time. It’s worth reading.

-------

1/ The workflow (quick summary)

This is the system that works:

- Plan clearly (PRD) @ChatGPTapp
- Break it down with TaskMaster
- Execute one task at a time in @cursor_ai
- Review everything with @coderabbitai

Let’s go deeper ↓
Jul 7 16 tweets 4 min read
If you’re using lovable to build apps, read this first

This is everything I wish I knew before starting ↓ Image Reposting this thread because many of you didn’t catch it the first time. It’s worth reading.

-------

1. Nail your first prompt

I always start inside my custom GPT, SnapPrompt, and get the full prompt for my landing page first.

This includes layout, structure, typography, and design style. I just copy-paste that into Lovable with a design reference attached and it gives me a clean starting point.
Jul 5 13 tweets 2 min read
I’ve used Claude Code for 100+ hours building MVPs.

Most developers are missing the best features.

Here are 10 tricks that will change how you build ↓ 1/ Use planning mode before EVERYTHING

- hit Shift+Tab twice to enter planning mode
- I plan every feature here first
- no code gets written until the plan is solid
- saves me hours of refactoring later
Jun 27 11 tweets 3 min read
This is how I made Cursor 10x more accurate.

GitMCP turns any GitHub repo into an MCP server.

Here’s how ↓ 1/ Cursor breaks when the context is messy

When you link a GitHub repo, Cursor loads everything even the junk.

That overloads the model.

And you get:

- Confused completions
- Useless suggestions
- Wrong assumptions
Jun 26 13 tweets 3 min read
I’ve vibe coded 18+ MVPs for clients using Cursor.

Security was the one lesson I learned the hard way.

Here’s the checklist I wish I had from day one ↓ 1. Rate limit your endpoints

If you skip this, bots or bad actors can hit your backend 100s of times per second.

This can:
• Crash your database
• Drain your Supabase usage
• Spike costs or open you to attacks

Tools to use:
• Supabase Edge Functions with a rate limiter
• Vercel Middleware
• Basic IP throttling with Next.js middleware
Jun 24 11 tweets 2 min read
Cursor just got 100x more powerful.

And it wasn’t because I used a bigger model.

It was because I gave it better memory.

Here’s how I fixed it with Context 7 MCP ↓ 1. The core problem: outdated context kills your agent

Every LLM has a training cutoff.

Ask it to install expo and it runs a broken command.

Not because the model is bad. It just doesn’t know recent changes.
Jun 23 12 tweets 2 min read
What I learned after building 2 MVPs with Claude Code (and why I might not go back to Cursor)

Here’s what I wish someone told me earlier ↓ 1/ Claude isn’t a one-shot coder

Don’t just ask it to “add feature X.”

Start in plan mode.

Let it think first.

Then review the plan and run it.

It’s slower, but the code is way cleaner.
Jun 19 12 tweets 3 min read
How I Make Cursor Write Perfect Code Every Time

This thread breaks down how I generate rules that make Cursor code cleaner, faster, and way more accurate ↓ TLDR:

• Set Global Rules for coding style
• Generate Project Rules using /generate cursor rules
• Use 5–6 custom .mdc files per MVP
• Attach rules as “Always”
• Let Cursor code with real context
Jun 18 12 tweets 3 min read
The workflow I wish I had when I started building MVPs

I use this to plan, design, build, and ship full MVPs in 2-3 weeks.

After 18 builds, here’s the exact system that works.

Bookmark this ↓ 1/ Plan with ChatGPT or Gemini

Before diving into the code, fully grasp the idea.

This phase is about achieving clarity. By its end, you should understand every aspect of the idea: the tech stack, core features, target audience, EVERYTHING.

Use ChatGPT or Gemini to create these docs:
- PRD/ MVP Plan
- UI Development Plan
- Database Design
- Implementation Plan
- Launch Checklist (you can generate this later, but use the same conversation)

Spend time here; it saves you from chaos later.
Jun 17 14 tweets 3 min read
This is the most efficient Cursor workflow I’ve used.

If you’re building MVPs fast, this combo is unbeatable:

TaskMaster + Cursor + CodeRabbit

Bookmark this. 1/ Why you need TaskMaster

Cursor is powerful, but as your MVP grows, you’re juggling:

• Feature backlog
• Refactors
• Bugs
• AI forgetting context

TaskMaster fixes that.

It acts like a project manager inside your terminal.

GitHub Link: github.com/eyaltoledano/c…
Jun 16 10 tweets 2 min read
Want to 10x your Cursor workflow?
Stop prompting like a developer.
Start thinking like a product manager.
Here’s how ↓ 1. AI tools are developers, not product managers

Cursor, Windsurf, Lovable, they’re great at execution.

But they can’t make product decisions for you.

If your prompt lacks clarity or structure, the output will suck. Every single time.
Jun 15 15 tweets 4 min read
If you’re using lovable to build apps, read this first
This is everything I wish I knew before starting ↓ Image 1. Nail your first prompt

I always start inside my custom GPT, SnapPrompt, and get the full prompt for my landing page first.

This includes layout, structure, typography, and design style. I just copy-paste that into Lovable with a design reference attached and it gives me a clean starting point.
Jun 11 12 tweets 3 min read
I ditched “vibe coding.”

And suddenly, building with Cursor felt calm again.

Here’s how I ship faster without burning out ↓ 1/ The workflow (quick summary)

This is the system that works:

- Plan clearly (PRD) @ChatGPTapp
- Break it down with TaskMaster
- Execute one task at a time in @cursor_ai
- Review everything with @coderabbitai

Let’s go deeper ↓
Jun 6 12 tweets 3 min read
I built 18+ MVPs, scaled my agency to $20K/month, and got acquired.

This is the exact AI dev workflow that made it happen ↓ 1/ The Problem With Traditional Agencies

Traditional dev agencies:
- Spend weeks on UI/UX before writing a single line of code
- Operate with large teams, slowing down decision-making
- Have high costs before validating an idea

The modern dev agency:
- Skips lengthy design phases and builds while iterating
- Uses small, efficient teams powered by AI
- Optimizes costs so founders can validate first
- Launches MVPs in under 4 weeks

I started @ignytlabs as a solo founder and delivered my first 6–7 MVPs alone.

Only after proving the model did I bring in other developers without bloated processes.
Jun 3 10 tweets 2 min read
How I use @lovable_dev + @cursor_ai to build MVPs 10x faster

I’ve used this exact combo to ship 18+ real products. Here’s how it works ↓ 1/ Start in Lovable

Every MVP starts inside Lovable.

Why? It lets me build full apps visually with:
• Supabase auth (email, Google, magic links)
• Stripe integration (checkout, subscriptions)
• OpenAI-powered features
• Fully responsive frontend flows

This saves me days of manual coding.
Jun 2 11 tweets 3 min read
I’ve vibe coded 18+ MVPs for clients using Cursor.
Security was the one lesson I learned the hard way.

Here’s the checklist I wish I had from day one ↓ 1. Rate limit your endpoints

If you skip this, bots or bad actors can hit your backend 100s of times per second.

This can:
• Crash your database
• Drain your Supabase usage
• Spike costs or open you to attacks

Tools to use:
• Supabase Edge Functions with a rate limiter
• Vercel Middleware
• Basic IP throttling with Next.js middleware
May 28 12 tweets 3 min read
If you’re not using /generate cursor rules, you’re leaving 80% of Cursor’s power on the table.

Here’s how to actually make it code like a real dev: 👇 1/ Why Cursor Rules Matter

Most people think Cursor just writes code for you.

But the truth is: the AI only works well when you train it right.

Cursor Rules are how you do that.

They tell the AI how to behave, what your project is about, and how you like your code written.
May 26 8 tweets 2 min read
Every AI Tool I Use to Build MVPs FAST for My Clients

As someone running an AI-first dev agency, speed is everything.

Here’s my go-to AI stack that has revolutionized how I build MVPs: 1/ @boltdotnew

I now start all my builds inside Bolt.

It lets me:
• Build full UIs with routing and logic
• Sync with Supabase for backend
• Integrate Stripe, FAST

It’s the fastest way to scaffold a functional MVP UI.

Example:
Built a subscription web app with auth, payments, and dashboards in under an hour. It would’ve taken days manually.