DEV Community: Shivansh

i turned a folder of markdown into a coworker that actually knows me

Shivansh — Tue, 02 Jun 2026 19:59:04 +0000

most people's notes are where ideas go to die.

a notion page you never open again. a notes app with 400 untitled entries. screenshots you saved and forgot. i had all of that.

so i built a second brain. it's a folder of plain markdown files that an AI reads, writes to, and thinks with. no fancy app. no subscription. just files and a model that knows the whole folder.

here's why i actually use it, and how it works.

why it's useful (the honest part)

the real reason this exists: i was tired of re-explaining myself to a chatbot every single time. who i am, what i'm working on, what i shipped last week. paste, paste, paste, then ask the question.

now there's one folder that holds all of it. work, projects, personal goals, raw notes. and the AI i use can read the whole thing at once.

so when i ask it something, it already knows the context. it knows i intern at remotestar. it knows kiks studios is my thing. it knows what moved this week and what's gone quiet. i don't feed it any of that. it's just there.

that's the unlock. context, already loaded, every time.

it goes both ways

this is the part people miss when they build a notes vault. most setups are read-only in practice. you write into them and never get anything back out.

mine is two-directional. i talk to it, it writes back into the files.

when i run a check-in, it asks me one question per work stream and takes my answers and edits the actual markdown. i'm not sitting there formatting notes by hand. i answer like i'm talking to someone, and it files everything where it belongs. later it reads that same stuff back to help me plan.

info goes in by conversation. it gets sorted to the right file. it comes back out as context when i need it. that loop is the whole thing.

how it's built

every note flows through the same lifecycle. capture it, sort it, distill it, then use it. tiago forte calls this CODE.

raw stuff lands in Inbox/ first. always. i don't decide where it goes when i capture it, because that's friction and friction kills the habit.

the folders aren't by topic. they're by how actionable the thing is. this is PARA. a project has a finish line. an area is ongoing, like my remotestar role. resources are reference i might pull from later. archive is done stuff i never delete.

everything is linked

here's where it gets good. every file links to other files. and shivansh.md, my master profile, sits in the middle as the hub everything connects back to.

my obsidian graph view. shivansh.md is the big node in the center, everything links back to it

that web isn't decoration. the AI follows those links. ask about one project and it pulls in the people, the client, the related notes, the personal context around it, because they're all connected.

this is why my planning actually works. when it gives me a weekly review or a morning brief, it isn't guessing. it walks the graph. one true profile in the center, everything else hanging off it, all reachable in one read.

the commands i run on top

a folder of notes is just a folder. the AI layer is what makes it feel alive. i wrote a few custom commands, each one reads the vault and does one job:

check-in walks me through each work stream and writes my answers back into the files. this is the two-way loop in action.
think reads everything and tells me what it notices. contradictions, a stream going quiet, leverage i'm missing. it doesn't ask questions or change anything, it just gives me an honest read.
morning hands me today's focus and a standup draft.
plan does a friday review and picks one focus per stream for next week.
spark turns what i'm actually working on into post ideas i could write today.

think and check-in are the two i lean on most. one keeps the vault current, the other tells me what i'm not seeing.

the voice rule

quick one. AI writing has a tell. em-dashes everywhere, "it's not x, it's y," lists of three, words like delve and leverage and robust.

so anything that comes out of this folder runs through a humanizer skill first. i imported it from github and tuned it to a few sample posts of mine, so the output reads like me typing instead of a press release. this blog went through it.

how it lives on two machines

i work off a macbook air and a mac mini. the vault sits in iCloud drive and both machines point a symlink at it. so does the config that holds the commands and the memory.

open obsidian on either mac and it's the same brain. start the AI on either one and it has the same context. nothing to sync by hand.

the takeaway

the value was never the notes. plenty of people have notes. the value is one place that's true, linked together, that a model can read all at once and write back into.

a markdown folder costs nothing. the linking and the two-way loop are what turn it from a graveyard into a coworker. build the folder, keep it honest, connect everything, then point a model at it.

your notes already know you. they just couldn't talk back yet.

Making GPay Smarter with Tags: UX for Real-World Spending

Shivansh — Sat, 19 Apr 2025 13:57:11 +0000

Introduction

Hi, I’m Shivansh, a multidisciplinary designer with a focus on UI/UX, graphic design, logos, product, and brand identity. I’m passionate about building meaningful digital experiences that help brands connect, communicate, and grow.

Here in this project I explore a new conceptual feature—Tags—for one of India’s most widely used payment platforms, Google Pay (GPay). The feature aims to improve how users track and organize their transactions through a smart and intuitive tagging system.

Disclaimer: This is a self-initiated, personal project. It is not affiliated with or commissioned by Google.

What is GPay?
Google Pay is a secure, user-friendly digital payments platform developed by Google and built on India’s UPI system. It enables seamless bank transfers, bill payments, and purchases—while ensuring top-tier encryption and integration across the Google ecosystem.

Why Google Pay?
I chose GPay simply because it’s the app I use every day—just like most people around me. It’s quick, clean, and does exactly what I need without getting in the way. That made it the perfect place to imagine a feature like Tags—something small and helpful that fits naturally into how we already use the app.

Project Overview

The project centers on integrating a new feature within GPay—Tags—designed to help users group, label, and track payments with more context. Whether it’s a road trip, a birthday dinner, or team expenses, Tags allow users to organize transactions under a single label, invite participants, and set budgets—all without disrupting GPay’s existing flow.

This case study dives into the challenges users face with managing scattered transactions and unclear payment histories—especially around group spending. Through structured UX methods like user personas, pain point mapping, wireframing, and visual design, the feature was crafted to solve these gaps while aligning with GPay’s overall simplicity.

Beyond UX, the solution also unlocks potential business value: better user retention, increased engagement, and more personalized financial insights—all within a familiar interface.

Role
UI/UX Designer
Tool Used: Figma

Problem & Goals

The Problem:
While GPay excels at quick transactions, it lacks structure for group or purpose-driven payments. Users struggle with:

No way to group related payments (like for trips or events).
Cluttered transaction feeds with no context or clarity.
Lack of tools for budgeting or tracking shared expenses.

This leads to confusion, especially when splitting bills or managing recurring group payments. Even you would have gone through it, sitting holding your head at the month end trying to figure out which payment went where.

The Solution:
Tags offer a lightweight but powerful way to organize payments. They aim to:

Group related transactions under custom labels (like “Goa Trip” or “Flat Rent”).
Add participants to shared tags for visibility and transparency.
Set simple budgets for events or time-bound tags.
Enhance GPay’s value without changing how users currently use it.

In short, Tags add a contextual layer that brings clarity to spending, without bloating the app’s core experience.

Design Process and Flow

1. Research & Discovery
It all began with a simple question—why do group payments still feel so messy? From weekend getaways to household grocery expense. It was very difficult to maintain track of expense went toward same cause. To dig deeper, I looked into how people currently manage shared expenses and explored tools like notes app (can you actually believe me). Patterns began to emerge, not just in how users struggled, but also in what they wished for. That’s when I started mapping goals—not just from a user’s point of view.

2. User Insights

I crafted user personas based on real-world use cases to build empathy. This phase helped define clear goals—what users needed, what frustrated them, and how Tags could make a difference in real usage.

Meet Rahul
A 26-year-old marketing executive living in Bangalore. He often pays bills be it Weekend trips or Friday dinners. Rahul uses Gpay daily but constantly struggles to track which payment was paid for what. Most of the time, the transaction history feed gets messy, and he ends up holding is head looking at 100s of payments trying to figure out. A feature like Tags would help him stay organized without switching to another app.

Meet Tanvi
A 30-year-old freelance designer who handles both personal and work payments on GPay. From paying vendors to managing client expenses, her transaction history is filled with unrelated entries. She wants an easy way to sort payments by project or event so she can stay on top of her finances. Tags would give her the control she needs—without changing how she already uses the app.

3. Wireframing & Ideation

I started by developing wireframes, with the goal of seamlessly integrating two new buttons—Create Tag and View Tags—onto GPay’s home screen. Placing them on the main screen was intentional. The idea was to make the feature feel like a natural part of the app, encouraging everyday use and making tag-based payments more accessible from the start.

Next was the task of creating a tag screen the entry point for tag ecosystem for users, This screen was designed to let users quickly create a Tag without feeling overwhelmed. I kept the layout minimal—just a name, optional budget, and participant field—to reduce friction and keep it lightweight. The rounded “Create Track” button draws focus, and recent tags at the bottom encourage reuse and familiarity.

My goal was to make this feel like a natural extension of GPay, not a new tool—so it blends with the existing UI while subtly introducing utility.

Next was my tags screen. this shows all active tags in one place—like a dashboard for past spending. Each tag neatly displays the total amount, number of transactions, and when it started. It’s designed to feel familiar, almost like viewing a folder of grouped expenses. The goal here was quick recall and a clean overview without digging through your feed.

Next, I designed the flow where users can add a tag to a transaction—either while making the payment or afterward. The idea was to keep it next to the existing “Note” field for familiarity.

Now about Notes: they’re useful for one-off context, but they fall short when payments start piling up. Notes don’t group transactions or help track shared spending over time. That’s where Tags step in—to bring structure, recall, and meaning when things scale.

View Tags screen. this screen shows the dedicated view for a tag. Once a tag is created, all related transactions show up here—sorted and grouped for clarity. It's a cleaner way to revisit your spending, whether for a trip, rent, or any category . Instead of scrolling through a long payment history, everything tied to that purpose lives here—simple and organized.

Final Results

Final Thoughts

Tags was my attempt to bring a little more sense and structure to everyday payments. It’s simple, contextual, and fits right into how people already use GPay—without adding friction.
A small idea, but one that could make payment tracking a lot more human.
— Shivansh