itszeang

I almost became a web developer. Then I realized AI would eventually do everything I was learning. So I pivoted. Here's what I'm doing instead:

@delveroin Citations. Not guesses.

pitch me your app in three words

@trying_to_exits claustrophobia

what's stopping u from coding here ?

@MicroLaunchHQ Building a local-first AI tool that turns your PDFs into structured, citation-backed outputs — summaries, arguments, and literature reviews — all running on your own machine.

What are you building this week?

@icanvardar At my internship, I take Excel files and turn them into usable workflows inside Workiom (a no-code ERP platform), then use Copilot to handle the data visualization part.

never met someone using copilot on this platform

@shub0414 You don’t need a completely new idea. You just need to take something that already exists and make it simpler, clearer, and genuinely useful for a specific type of person. People don’t pick what’s new, they pick what actually works for them.

Every idea feels taken. Every API already exists. Every SaaS has 12 competitors. So what do we even build now?

If you drop your PDFs in, it turns them into a private, citation-backed research workspace. Ask anything → answers come only from your documents Generates structured summaries, arguments, and literature reviews Every claim is traceable to a source (no hallucinated citations) Works offline — your data never leaves your machine Export ready-to-use text → .docx with references It’s basically: your own papers → organized thinking → publishable output

I’ll pay for your SaaS today 💰 No landing page. No waitlist. No buzzwords. Just tell me what it does 👇

2024 : prompt eng 2025 : vibe coder 2026 : ruler of agent 2027 : ??? (unemployed)

Built in: → 1 day for initial RAG system → 2 hours to add local LLM → 15 PDFs indexed (my course notes) Code is open source: github.com/itszeang/local… Building in public. Learning AI engineering. Follow for more 🚀

Demo: Query: "What is Daily Scrum?" The system: → Retrieves 3 most relevant chunks → Shows relevance scores → Generates natural language answer All locally. All private.

I just built a fully LOCAL RAG system. No OpenAI API. No cloud. No costs. 100% privacy. 100% control. Here's how it works 🧵

@ryankatayi @realmfberlin idk why but his gives a ridiculous motivation

@itszeang @realmfberlin 😂😂😂what are you even on about?

Fastest way to go broke?

@stuartpididiot @agNavya 400iq

@itszeang @agNavya If listening is faster than reading to you then you read slow as fuck holy shit

Is it humanly possible to read 200 pages in a single day?

@Tartopoile @agNavya Do I really have to tell you not to do this every time?

@itszeang @agNavya wtf no ? did you ever read something were you've been hooked up and can't stop ?

@oluwvtosin @realmfberlin You needed it (it was definitely needed).

@itszeang @realmfberlin This is me 🫩

@Anas_founder The building process is the most exciting part for me compared to the others. I enjoy it the most because that's when I’m actually learning and improving myself.

As a founder, what do you enjoy most? 1- Building 2- Marketing 3- Talking to users

@feifei_qiu Create a brutally honest custom Gem or GPT, or whatever you use.

Where do you get honest feedback? Friends won’t hurt you. Strangers won’t care.