Skip to content

Daksh Bathla

Builder/Developer/Founder

I build AI products that solve real problems, one at a time.

I don't read about building.
I build.

I'm an 18-year-old computer applications student from Delhi. Most of what I know came from shipping things that were slightly too hard, then fixing them in public.

Navital Labs is the studio the serious products ship under. Everything else is an experiment, and I keep the line between them obvious.

how-i-work.ts
// every product I have shipped started hereconst idea = notice(somethingAnnoying) while (!usedByRealPeople(idea)) {  build(idea)  ship(idea)  // before it feels ready  idea = learnFrom(whatBreaks)}
git logmain
  • a91f3c2feat(leadscrapper): score prospects by digital presence
  • 4d0b7eafix(itiner): keep agent edits when the AI redrafts
  • c17e55bfeat(canvas-zero): open the auction house
  • 9fbb204refactor(nucleus): split the grid layer out
  • 2e6a180chore(navital): one studio, one standard

leadscrapper.pro

live

itiner.in

live

canvas-zero

building

Products, assembled.

Every one of these started as a wireframe and an argument with myself about whether it was worth building. Here is where they ended up.

LeadScrapper Pro interface

01Live/ Navital Labs

LeadScrapper Pro

AI prospecting software for freelancers and agencies.

The hard part

Most prospecting tools either find businesses or write the outreach. Putting discovery, digital-presence analysis and messaging in one pass meant building a pipeline where the signal scraped at the start still survives, intact and attributable, by the time a draft is written.

  • Next.js
  • Node.js
  • PostgreSQL
  • Web Scraping
  • AI
Itiner interface

02Live/ Navital Labs

Itiner

AI-assisted itinerary software for travel agencies.

The hard part

Travel agents do not want a model overwriting their judgement. The hard part was a drafting flow where the AI can regenerate an itinerary without ever destroying the edits a human has already made — and where the client-facing output still looks like the agency, not like a chatbot.

  • Next.js
  • TypeScript
  • PostgreSQL
  • AI
  • Tailwind
Canvas Zero interface

03Building

Canvas Zero

A browser game about painting, auctions, and creative pressure.

The hard part

A painting sim is a toy until choices cost something. Wiring the auction loop back into the studio — so what you make, and when you stop making it, actually matters — is what turned a drawing surface into a game.

  • React
  • Next.js
  • Canvas API
  • TypeScript

How it went.

  1. Started building for the web

    Learning by shipping instead of by tutorial. HTML, CSS, and the habit of putting things online before they were finished.

  2. Nucleus CSS

    A lightweight, mobile-first CSS framework released as open source — the first thing I built for other developers rather than for myself.

    Open ↗
  3. Last Minute AI

    AI study help for students working under exam pressure. Revision workflows, summarised material, and flashcards.

    Open ↗
  4. Canvas Zero

    A browser game about painting, auctions, and creative pressure. Built to find out how much product a single canvas element can hold.

    Open ↗
  5. Itiner

    A hackathon prototype that turned into a B2B product: AI drafts itineraries, travel agents refine and publish them.

    Open ↗
  6. LeadScrapper Pro

    AI prospecting software for freelancers and agencies. Business discovery, digital-presence analysis, and outreach in one workflow.

    Open ↗
  7. Founded Navital Labs

    The studio the serious products ship under. LeadScrapper Pro and Itiner live here; experiments stay clearly separate.

    Open →
  8. Now

    Building AI-native products, writing about the work, and studying what interfaces become when AI is the primary actor.

    Open →

Not a list. A system.

Nothing here is useful alone. Hover a discipline to see what it is wired into.

Default view

Backend

  • Node.js
  • PostgreSQL
  • Go
  • REST APIs

wired to Frontend, AI, Cloud, Automation

I build products.

I build experiences.

I build things people actually use.

Let's build something.

Open to product conversations, collaborations, and ideas that should exist but don't yet. The terminal works — type in it.

daksh@dakshbathla.in — zsh

type 'help' to begin