DEV Community: Sai Shravan Vadla

Check out Chatly

Sai Shravan Vadla — Sat, 06 Jun 2026 05:57:47 +0000

GitHub “Finish-Up-A-Thon” Challenge Submission

Sai Shravan Vadla

Jun 6

I Finally Finished Chatly: From Unfinished Project to Cross-Platform App

#devchallenge #githubchallenge

Comments

5 min read

I Finally Finished Chatly: From Unfinished Project to Cross-Platform App

Sai Shravan Vadla — Sat, 06 Jun 2026 05:56:19 +0000

This is a submission for the GitHub Finish-Up-A-Thon Challenge

What I Built

Chatly is a modern real-time chat platform designed to be more than just messaging. Chatly combines real-time conversations, AI-powered assistance, user profiles, paywall-protected premium features, dark mode, push notifications—all in one sleek, SaaS-ready package.

It started as a project I built while learning modern web development and AI integrations. Over time it accumulated features, but it never quite felt finished. This challenge gave me the excuse to come back, polish the rough edges, build a native Android app, and finally ship something I was proud to share.

For this challenge, I decided to revisit Chatly and finally give it the attention it deserved.

The result is a significantly more polished product with improvements across the web experience and a new native mobile experience built with Expo and React Native.

Some highlights include:

User authentication
Persistent chat history
Responsive web application
Native Android application
AI-powered conversations
Message attachments including images, videos, audio and files
Web push and Expo notifications
Improved user experience and visual polish
Better overall reliability and usability

More importantly, Chatly has now evolved from a learning project into a real application that people can use.

Demo

Please read "Important info regarding Chatly's limitations" below before testing the demo.

Web App

https://chatly-brown.vercel.app
(An account is needed to use Chatly. Email or Google/Github authentication is supported. If using email, an email verification is required.)

Android Download

https://chatly-brown.vercel.app/download
(Due to limitations, Chatly is only available as an app on Android as a .apk file downloadable from Expo. This file is secure. You might need to allow "install unknown apps" in your browser to install Chatly on Android.)

GitHub repository

shravzzv / chatly

Chatly is a modern real-time chat platform designed to be more than just messaging. Chatly combines real-time conversations, AI-powered assistance, user profiles, paywall-protected premium features, dark mode, push notifications— all in one sleek, SaaS-ready package.

Chatly

Chatly is a real-time 1:1 messaging application with media attachments, AI-assisted message enhancements and usage-based billing.

The project is structured as a cross-platform monorepo supporting both web and native clients backed by a shared Supabase infrastructure.

Chatly is designed as a production-grade SaaS example covering authentication, messaging, media uploads, rate limiting, billing, and enforcement — without placeholder flows.

Features

1:1 Text Chat: Real-time messaging with delivery states.
Media Attachments: Support for images and file sharing.
AI Enhancement: Integrated AI-assisted message refinement.
Usage Enforcement: Daily limits and plan-based feature gating.
Real-time Infrastructure: Powered by Supabase Realtime.
Secure Auth: Email and Google OAuth integration via Supabase Auth.
Subscription Billing: Full Lemon Squeezy integration including a customer billing portal.

Tech Stack

Web

Next.js (App Router)
React
TypeScript
Tailwind CSS

Mobile

Expo
React Native
TypeScript
Expo Router

Backend

Supabase (Postgres, Auth, Realtime, Storage)

Infrastructure

Vercel AI…

View on GitHub

Screenshots

The landing page:

Signup page:

Pricing page:

Chat dashboard:

Plan page with usage:

Account page:

Video Walkthroughs

Web

Android

Important info regarding Chatly's limitations

Chatly uses LemonSqueezy as its Merchant of Record. The store is currently awaiting final approval, which means the LemonSqueezy billing portal is not yet available. To allow the subscription experience to be evaluated, Chatly remains connected to LemonSqueezy's test payment environment. All subscription-related functionality can therefore be explored safely without real charges. Once store approval is completed, billing management will be available through LemonSqueezy's standard customer portal. Nevertheless, Chatly is completely usable even with this limiation.

All of Chatly's subscription plans are available at https://chatly-store.lemonsqueezy.com/checkout when navigated using the links within the app and these can be purchased using test data.

Use test card details such as 4242 4242 4242 4242 with any valid future date for expiry and any 3 digit number as the CVC. The rest of the details aren't crucial for purchasing a subscription.

The only real limiation here is that LemonSqueezy billing portal (https://chatly-store.lemonsqueezy.com/billing) is unavailable because of the store not being activated. Due to this, managing billing is not possible at the moment. Once activated, users can manage their plans here. But until then, test data can be used and plans can be explored as if real transactions were made. If multiple subscriptions are present, only the highest value one gets used.

The Comeback Story

Before	After
Web application only	Web + Android
Basic chat functionality	Chat + attachments + enhancements
Limited settings	User settings and personalization
Prototype	Publicly deployed product
Unfinished side project	Actively maintained application

This release contains all the changes I've made for phase 2 of this project. It contains all the work completed for the Finish-Up-A-Thon challenge and represents the transformation of Chatly from an unfinished side project into a cross-platform application.

Before

When I originally worked on Chatly, my focus was primarily on getting functionality working.

The application could already support basic chat conversations, but several parts of the overall experience were incomplete:

Mobile experience was limited
User experience needed refinement
Various UI and workflow improvements were still unfinished
The project lacked the level of polish needed for a public launch
There was no dedicated native mobile application

Like many side projects, it became something I would occasionally revisit but never fully complete.

After

For the Finish-Up-A-Thon, I decided to treat Chatly like a real product and finish what I started.

Major improvements included:

Native Mobile App

The biggest addition was building a native mobile version using Expo and React Native.

This transformed Chatly from a web-focused application into a cross-platform experience.

Web App Polish

I revisited numerous areas of the web application and improved the overall experience:

Better interface consistency
Improved workflows
Bug fixes
General usability improvements
Additional polish throughout the application

Product Readiness

I also spent time improving the parts of software development that users rarely see but always feel:

Reliability improvements
Better overall user experience
Cleaner codebase
Production-focused refinements

The project now feels substantially closer to the vision I originally had when I started building it.

My Experience with GitHub Copilot

GitHub Copilot played an important role in helping me finish the project.

I primarily used Copilot as a development partner while implementing features, refactoring existing code, and solving implementation problems.

Some of the ways Copilot helped included:

Generating repetitive boilerplate code
Assisting with React and React Native development
Accelerating component creation
Helping troubleshoot implementation issues
Speeding up refactors
Providing alternative implementation ideas

One of the biggest benefits was maintaining momentum. Instead of getting stuck on smaller implementation details, I could stay focused on improving the product itself.

For a challenge centered around finally finishing a project, that productivity boost was incredibly valuable.

What's Next?

Chatly is no longer an abandoned side project.

My goal now is to continue improving the application, expand platform support, and keep refining the user experience based on real user feedback.

Finishing the project reminded me that shipping and polishing often matter more than starting something new.

Sometimes the best project idea is the one you've already started.

This week's ship: Expensely 🪙

Sai Shravan Vadla — Tue, 26 Aug 2025 16:54:33 +0000

Introduction

Hey everyone 👋,
I’m back again with another project for my “ship something every week” challenge! This week’s build is Expensely — a clean and minimal income/expenses tracker.

👉 Live demo: expensely-money.vercel.app
💻 GitHub repo: github.com/shravzzv/expensely

shravzzv / expensely

Expensely is a simple finance tracker designed to make money management less overwhelming. Add income, record expenses, and view your balance in real time. Expensely gives you the essential insights you need without unnecessary complexity.

Expensely

Project demo

Description

Expensely is a simple, clean, and minimalistic expense tracker built with Next.js, Tailwind CSS, and ShadCN.
It helps you keep track of your income, expenses, and overall net balance in a clear dashboard view.

Stay on top of your money with an elegant and lightweight tool.

Features

📊 Add and manage income & expenses
🧮 Automatic net worth calculation
🎨 Clean UI with ShadCN components
⚡ Fast and responsive (built on Next.js & Tailwind CSS)
☁️ Deployed on Vercel for instant access

Technologies Used

Installation

To install the project, follow these steps:

git clone https://github.com/shravzzv/expensely.git
cd expensely
npm install
npm run dev

How to Contribute

If you'd like to contribute, follow these steps:

Fork the repository on GitHub.

Clone your fork locally.

git clone [your_forked_repository_link]
cd expensely

Create a new branch for your feature or bug fix.
```
git checkout -b feature-or-bug-fix-name
```
Make your changes, commit…

View on GitHub

💡 Why I Built Expensely

Money is one of those things we all think about every single day — but tracking it often feels like a chore.

I’ve tried using spreadsheets and complicated finance apps in the past, but they were either too clunky or too bloated with features I didn’t need.

So I thought: What if I could make a lightweight app that just lets me quickly log my income and expenses, and shows me my net worth instantly?

That’s how Expensely was born.

⚙️ What’s Inside

Here’s what I built into Expensely this week:

✅ Add new transactions (income or expense)
✅ Edit or delete transactions
✅ See your total income, total expenses, and net worth at a glance
✅ Clean, responsive UI built with Next.js, TailwindCSS, and ShadCN
✅ Instant feedback with toast notifications

It’s simple, fast, and distraction-free — perfect for keeping track of your money without overwhelm.

🚀 Tech Stack

I kept things minimal but modern:

Next.js — frontend framework
TailwindCSS — styling
ShadCN — beautiful UI components
React state for managing transactions (no backend yet — everything runs locally in the browser)

📈 What’s Next?

Right now, Expensely is just a frontend app. But here are a few ideas I’m considering:

Add authentication & databases (maybe Supabase or Firebase)
Categories & filtering for expenses
Exporting to CSV/Excel for easier budgeting

🙌 I’d Love Your Feedback

If you give Expensely a try, I’d love to hear:

Does the simple approach work for you, or would you prefer more advanced features?
What’s the one must-have feature you’d add?
Would you actually use this as your daily money tracker?

👉 Try it here: expensely-money.vercel.app
👉 Star the repo if you like it: github.com/shravzzv/expensely

Thanks for reading — and see you next week with another small project 🚀

— Sai

Introducing Pomora: a minimal, focused Pomodoro timer (Next.js + Tailwind + shadcn/ui)

Sai Shravan Vadla — Tue, 19 Aug 2025 09:04:04 +0000

Introduction

“Do one thing well.” That’s the entire idea behind Pomora. It’s a simple Pomodoro timer with a clean UI, dark mode, and zero fluff—built with Next.js (App Router), TypeScript, Tailwind CSS, and shadcn/ui.

Live demo: https://pomora.vercel.app/
Source code: https://github.com/shravzzv/pomora

Why build another Pomodoro timer?

Two reasons:

Build-every-week habit: I wanted a small but polished app to sharpen my Next.js + UI chops.
Minimalism: Most timers do too much. I wanted a calm, unobtrusive tool I actually enjoy keeping open.

No accounts. No dashboards. No backend. Just a good timer.

Features

Three modes: Pomodoro (25m), Short Break (5m), Long Break (20m)
Controls: Start / Pause / Resume / Restart + Reset
Smart time display: hh:mm:ss only when hours are needed
Dark/Light themes with a settings dialog
Responsive and accessible UI (buttons, focus, contrast)
Privacy friendly: no backend; optional Vercel Analytics if you want it

Tech stack

Next.js (App Router) + TypeScript
Tailwind CSS for utility-first styling
shadcn/ui for accessible primitives (Tabs, Dialog, Button, Switch)
lucide-react for crisp icons
Vercel Analytics

Wrap-up

Pomora is intentionally small. If you’ve been meaning to ship something, I highly recommend a scoped, well-designed utility like this. You’ll finish faster and learn a lot on polish.

If you try it, I’d love feedback.

🦺 CrisisCopilot – Your AI Agent for Personal Emergency Management

Sai Shravan Vadla — Fri, 04 Jul 2025 15:32:00 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

🧠 Problem

When you're hit with a personal emergency—like a medical crisis, travel delay, sudden job loss, or family issue—your brain goes into survival mode. It becomes difficult to think clearly, prioritize, or remember critical tasks. Often, people don’t know who to contact, what to do next, or how to coordinate their life while navigating a stressful situation.

There’s no single system that calmly takes over, executes key coordination steps, and lets you focus on what matters most: your well-being.

💡 Solution

CrisisCopilot is a Runner H agent that helps you handle personal emergencies with clarity and control. It automates coordination, communication, documentation, and scheduling across your connected tools.

Just upload your emergency contacts as a .csv file, describe your situation, and CrisisCopilot does the rest—communicating with your contacts, generating your action plan, setting calendar events, tracking expected expenses, and informing your team—all while you catch your breath.

What I Built

CrisisCopilot is a multi-step Runner H automation that helps you manage real-life emergencies without cognitive overload. It spans five tools (Gmail, Google Docs, Google Calendar, Google Sheets, and Slack) and uses one input (your emergency context + contacts CSV) to orchestrate a full support workflow.

I built it to simulate the experience of a real executive assistant who calmly says, "Don’t worry, I’ve got this." Except this assistant never sleeps.

Demo

Watch CrisisCopilot in action

Surfer 2: The Next Generation of Cross-Platform Computer-Use Agents - H Company

H Company builds models, agents, and products that automate tasks and simplify complex work. We empower people and enterprises to move faster, think bigger, and do more of what matters.

hcompany.ai

The Google doc CrisisCopilot has created

Fire Emergency – To-Do List - Google Docs

docs.google.com

How I Used Runner H

💼 Prompt Format Used in Runner H:

⚠ Note that you need to upload a .csv file of your emergency contacts first. You can get them by exporting them from Google contacts or any other contacts software as a .csv file. Then upload them to Runner H.

Then proceed to use this prompt below with all of the necessary connections--Gmail, Docs, Sheets, Drive, Calendar and Slack.

You are CrisisCopilot, my personal AI assistant for emergency management.

Emergency: There was a fire in my apartment building and I had to evacuate with just my phone and wallet. The fire department is here, but I don’t know how bad the damage is. I’m safe, but shaken. I need to find emergency shelter for tonight and figure out how to access insurance and document the damage. My laptop and important documents may be gone.

Your job is to calmly coordinate all my next steps across tools I've connected. Follow this workflow carefully:

---

1. **Emergency Contacts**
   - Read the uploaded CSV file titled `emergency_contacts.csv`
   - Send a respectful, concise Gmail message to each contact describing the emergency and letting them know I may reach out soon. Keep tone calm and reassuring.

   ✅ *Example Gmail message:*

Subject: Quick Update – \[Your Name] is handling a personal emergency

Hi \[Contact Name],

I wanted to let you know that I'm currently dealing with a personal emergency: \[Brief one-line description, e.g., "a minor accident and medical recovery"]. I'm okay, but things may be a little unpredictable over the next few days.

Just wanted to keep you in the loop — I’ll follow up soon if anything needs your help or input.

Warmly,
Sai Shravan Vadla

---

2. **Create a Google Doc**
- Title: "[Emergency Type] – To-Do List"
- Generate a list of 10–20 practical, organized to-dos to navigate this emergency.
- Categorize tasks by priority: *Immediate*, *This Week*, *Follow-up*
- Keep it clear, structured, and helpful.

---

3. **Add Google Calendar Events**
- Create 10-15 events in my calendar ranging the next 10-12 days about any todos you might think are necessary.

---

4. **Create Google Sheet: "Emergency Expense Tracker"**
- Columns: `Item/Service`, `Estimated Cost`, `Actual Cost`, `Date`, `Status`
- Add a few placeholder rows based on likely costs related to this emergency.
- Include a row at the bottom to show total estimated vs. actual costs.
This is really important. Please create this Google sheet at all costs. This should be a sheet and not a document.
---

5. **Send Slack Message to My Team**
- Post a thoughtful message to our general team Slack channel to keep everyone in the loop and reduce confusion.
- Message should be empathetic, not overly dramatic, and set clear expectations.

✅ *Example Slack message:*

Hey everyone — just a quick heads up: I'm currently handling a personal emergency (\[brief mention, e.g., "some health stuff"]). I'm okay, but I may be slower to respond or need to delegate some tasks over the next few days.

I’ll keep you posted if anything urgent changes. Thanks so much for your understanding and support 🙏

– Sai Shravan Vadla

---

6. **Final Summary PDF**
- Generate a PDF summary titled 'CrisisCopilot report':
  - What actions were taken, to whom were the emails sent to
  - What tools were updated
  - Link to the Google Doc, Sheet, Calendar events, and Slack message

Use Case & Impact

Who is this for?

Remote workers or freelancers with no backup
Students or digital nomads
Solo founders juggling work and life
Anyone who panics in a crisis or forgets critical steps

How does it help?

Streamlines chaos into a plan
Notifies your people automatically
Keeps your calendar and budget in sync
Gives you space to breathe and focus on healing

This is especially powerful for people living away from family or with executive function challenges (like ADHD), who need support in moments of stress.

Social Love

// Detect dark theme var iframe = document.getElementById('tweet-1941158637499736257-433'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1941158637499736257&theme=dark" }

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

linkedin.com

🙏 Final Thoughts

CrisisCopilot is just the beginning. In the future, it could:

Look up insurance policies
Connect with therapy services
Monitor medical documents or prescriptions
Add recurring wellness reminders

Runner H has made it incredibly easy to prototype something so real and so personal. I'm grateful to build something that could genuinely help someone in their toughest moments.

Thanks to H Company and DEV for this challenge.

📈 TrendSage – An AI Agent That Tracks Trends So You Don’t Have To

Sai Shravan Vadla — Tue, 01 Jul 2025 14:50:00 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

🌍 The Problem

The internet is a chaotic, noisy place. For founders, creators, marketers, and investors, keeping up with emerging trends across industries like AI, health tech, or the creator economy takes hours of skimming blogs, newsletters, Substacks, Twitter, and podcasts.

Even then, you risk missing critical shifts or wasting time on shallow content. There's no simple way to stay informed, organized, and proactive—without burning out.

🤖 The Solution: TrendSage

TrendSage is an AI agent that browses the web for you, discovers the latest trends in your chosen domain, and compiles them into shareable, organized content across multiple platforms.

It’s like having your own AI-powered market research assistant that delivers clarity without the noise.

What I Built

TrendSage is a Runner H agent that automates the entire market trend discovery and delivery workflow:

🔍 Web Research – It browses the internet and finds 3–5 real emerging trends in your chosen category.
📝 Google Doc Report – It generates a structured, professional market brief with summaries and examples.
📊 Google Sheets Table – It creates a table of key takeaways for quick scanning or internal use.
📩 Gmail Digest – It sends a summary of the findings to your inbox.
💬 Slack Message – It posts a quick summary into your Slack workspace for async team awareness.
🧾 PDF Recap – It wraps it all up in a downloadable PDF for sharing or archiving.

Demo

See TrendSage in action here:

Surfer 2: The Next Generation of Cross-Platform Computer-Use Agents - H Company

H Company builds models, agents, and products that automate tasks and simplify complex work. We empower people and enterprises to move faster, think bigger, and do more of what matters.

hcompany.ai

The Google Doc TrendSage created:

TrendSage Report – July 1, 2025 - Google Docs

docs.google.com

The Google Sheet TrendSage created:

TrendSage Summary – July 1, 2025 - Google Sheets

docs.google.com

How I Used Runner H

Tools Connected:

Google Drive (stores created files)
Google Docs (create formatted summary report)
Google Sheets (create formatted spreadsheet report)
Gmail (send follow-up emails to clients)
Slack (post updates to team)

Prompt

You are **TrendSage**, a market intelligence agent for entrepreneurs and product teams. Your job is to find, summarize, and organize emerging trends from the internet into an easy-to-understand weekly report. Follow these steps exactly:

---

### **Step-by-Step Instructions:**

1. **Search the Web for Trends**
   Browse the internet for **3–5 emerging trends** in one of these domains:

   * AI tools and products
   * Creator economy
   * Future of work
   * Climate tech
   * Education technology
   * Health tech
     Use only **publicly accessible content** from blogs, news sites, subreddits, Twitter, and forums.

2. **Extract Key Info for Each Trend**
   For each trend you find, include:

   * 🔹 **Trend Name**
   * 🧠 **What it is** (1–2 sentence summary)
   * 📈 **Why it matters** (impact, opportunity, etc.)
   * 🌍 **Source(s)** with link(s)

   > Example:
   > **Trend Name:** AI Co-Pilots for Education
   > **What it is:** Tools like MagicSchool and Khanmigo are using LLMs to help teachers with lesson planning, grading, and personalization.
   > **Why it matters:** These tools could massively reduce teacher workload and improve student outcomes.
   > **Sources:** [https://www.magicschool.ai/](https://www.magicschool.ai/)

3. **Write the Full Report in a Google Doc**
   Create a new Google Doc titled:
   ➤ `TrendSage Report – [Today's Date]`

   Format it professionally using:

   * A clear headline for each trend
   * Bullet points for “What it is” and “Why it matters”
   * Hyperlinked sources
   * Optional intro paragraph: “This week’s trend roundup...”

4. **Create a Summary Sheet**
   Create a **Google Sheet** titled:
   ➤ `TrendSage Summary – [Today's Date]`

   Use this column structure:
   \| Trend Name | Summary | Source URL(s) | Category | Date |
   Fill in each row with the trends you discovered.

5. **Send a Slack Message**
   Post a summary message to Slack:

   > 🚀 TrendSage Weekly Report is ready!
   > Found \[X] new trends in \[Chosen Category].
   > 📄 Full report: \[Google Doc link]
   > 📊 Sheet summary: \[Google Sheet link]

6. **Send a Gmail Digest (Optional)**
   Send an email to yourself or a specified email:
   **Subject:** TrendSage Weekly Report – \[Date]
   Include a short summary + links to the report and sheet.

7. **Generate a Final PDF Output**
   Create a PDF file titled `TrendSage Digest – [Date]` summarizing:

   * The 3–5 trends
   * Short bullet-point takeaways
   * Links to sources
   * Link to the full Google Doc report and Sheet

---

### 💡 Notes

* Focus on **fresh trends** that are less than 1 month old.
* Prioritize clarity and usefulness over length.
* If no good trends are found, return a polite message stating that nothing relevant was found and stop.

Use Case & Impact

Target users:

Startup founders looking for product trends
Marketers tracking industry movement
Writers/creators who need content inspiration
VCs and analysts who want faster deal sourcing insights

Impact:

Instead of spending 1–2 hours a day reading across tabs and newsletters, TrendSage gives users a focused brief they can use to make decisions or share with a team. It’s fast, shareable, and multi-platform.

Social Love

// Detect dark theme var iframe = document.getElementById('tweet-1940071634364506443-245'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1940071634364506443&theme=dark" }

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

linkedin.com

Thanks for checking out TrendSage!

📊 ClientPulse – Your AI Assistant for Client Follow-Ups & Summaries

Sai Shravan Vadla — Sat, 28 Jun 2025 05:18:37 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

Intro

Managing multiple clients means juggling meeting notes, follow-up tasks, important deadlines, and communication — often across scattered documents and inboxes.

It’s easy to forget a detail, miss a deadline, or delay an important check-in.

Solution

ClientPulse is an AI-powered agent built using Runner H that processes client meeting notes, tracks follow-ups, and handles outreach automatically.

Just upload your client notes — ClientPulse does the rest.

What I Built

ClientPulse is a multi-step automation that:

📥 Reads uploaded client meeting documents (.docx, .txt, .pdf, or Google Docs)
🧠 Extracts key details: client name, contact info, discussion summary, follow-ups, and deadlines
📄 Creates a clean Google Doc summary for all clients
📅 Schedules follow-up tasks as events in Google Calendar
📧 Sends a follow-up email to the client using Gmail (based on the notes)
💬 Posts a Slack summary to your internal channel
📎 Generates a PDF report summarizing everything the agent did

Demo

Watch ClientPulse in action:

Surfer 2: The Next Generation of Cross-Platform Computer-Use Agents - H Company

H Company builds models, agents, and products that automate tasks and simplify complex work. We empower people and enterprises to move faster, think bigger, and do more of what matters.

hcompany.ai

An example client meeting notes that was uploaded to the agent:

📄 1. ClientPulse – Acme Corp – Website Kickoff
Date: June 25, 2025
Client Contact: Sarah Mitchell – Head of Marketing
Email: sarah.mitchell@acmecorp.com

📝 Meeting Summary:
We kicked off the new website redesign project for Acme Corp. Sarah emphasized the need for a more modern, mobile-first design and faster performance. They're also rebranding, so the new design must align with their updated color palette and brand voice.

They would like a rough homepage wireframe by July 3 and expect a first prototype by July 12. SEO will be a key focus, so collaboration with their SEO team is essential starting next week.

✅ Action Items:
[You] Deliver homepage wireframe by July 3

[You] Begin competitor benchmarking (look at Stripe, Notion, and Intercom)

[Acme] Provide updated brand guidelines by June 28

[You + Acme SEO team] Meet to align SEO priorities – scheduled for July 1 at 10 AM

📌 Notes:
Their development team is using Webflow, so designs must be compatible.

Sarah mentioned their CEO might want a review session mid-July.

The Google Doc client report created by ClientPulse (Note: I initially tried to have Runner H create a Google Sheet for this, but due to consistent output issues, I opted for a well-structured Google Doc instead):

ClientPulse – Tracker Report - Google Docs

docs.google.com

How I Used Runner H

Tools Connected:

Google Drive (stores created files)
Google Docs (create formatted summary report)
Google Calendar (create follow-up events)
Gmail (send follow-up emails to clients)
Slack (post updates to team)

Prompt

You are ClientPulse, an AI assistant that helps consultants and freelancers manage client interactions.

1. Read all uploaded client notes. These could be .txt, .docx, .pdf, or Google Docs.

2. For each document, extract:
   - Client name
   - Email address
   - Meeting summary
   - Action items / follow-ups
   - Any mentioned deadlines or dates

3. Create a formatted Google Doc titled: "ClientPulse – [Date] Summary" with this structure:
# ClientPulse – [Date]
## Client: [Client Name]
**Email:** [email@example.com]  
**Meeting Summary:**  
[Concise summary here]  
**Follow-Ups:**  
- [Follow-up 1]  
- [Follow-up 2]  
**Deadlines:**  
- [Task] – [Date]

(Repeat the above section for each client.)

4. For each deadline mentioned, create a Google Calendar event with:
   - Title: "[Client Name] – [Task]"
   - Date: [Due Date]
   - Description: auto-filled from the follow-up info

5. Send a follow-up email to the client using Gmail:
   - To: [Client Email]
   - Subject: "Follow-Up from Our Meeting"
   - Body: Use the meeting summary and follow-up list as a personalized message

6. Post a Slack message summarizing all client actions:
   "📝 ClientPulse Summary: [Client A] – 2 follow-ups, 1 deadline. [Client B] – 1 follow-up, 0 deadlines. Full doc: [link]"

7. Export the final summary Google Doc to PDF and save it in Drive.

Make sure all information is clear, formatted, and concise.

Workflow

Upload client meeting notes
Runner H processes the files
Extracts key info per client
Creates Google Doc with all summaries
Schedules events in Google Calendar
Sends follow-up emails via Gmail
Posts Slack summary
Creates a final PDF report

Client Notes ➜ ClientPulse Agent
           ├──> Google Doc Summary
           ├──> Google Calendar Events
           ├──> Follow-Up Emails via Gmail
           ├──> Slack Update
           └──> PDF Summary in Google Drive

Use Case & Impact

Perfect for:

Freelancers or consultants managing multiple clients
Agencies with weekly client check-ins
Project managers handling external stakeholder notes

Impact:

Ensures no client detail is lost or forgotten
Saves hours of manual follow-up work
Keeps internal teams updated automatically
Strengthens professionalism with timely, personalized communication

Social Love

// Detect dark theme var iframe = document.getElementById('tweet-1938831985709043868-484'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1938831985709043868&theme=dark" }

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

linkedin.com

Conclusion

ClientPulse started as a way to automate messy client admin work — and it turned into a fully functional AI assistant that handles note processing, outreach, and reporting.

In the future, I’d love to explore:

CRM integration (like HubSpot or Notion)
Client satisfaction score prediction based on notes
Voice-to-note processing for call summaries

Thanks to the Runner H and DEV team — building this was a blast and incredibly practical.

🚀 StudySherpa: An AI Agent to Organize Notes, Generate Flashcards, and Track Deadlines

Sai Shravan Vadla — Fri, 27 Jun 2025 08:33:29 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

Intro

Being a student often means dealing with chaos:

Notes scattered across multiple docs
No system for review or spaced repetition
Deadlines scattered across syllabi and forgotten calendar invites

This is cognitively exhausting, and students end up spending more time organizing than learning.

Solution

StudySherpa is an AI-powered agent built with Runner H that solves this problem. It reads uploaded study materials, creates flashcards, adds deadlines to Google Calendar, and keeps everything centralized and actionable.

It’s like having a personal academic assistant that organizes your knowledge and schedule while you sleep.

What I Built

StudySherpa is a Runner H automation designed to:

Read docs containing study content or syllabus
Extract important points and generate Q&A-style flashcards
Write these flashcards into new Google Docs
Create Google Calendar events for any upcoming deadlines found in the uploaded documents
Create a Flashcard Index linking all generated flashcard docs for easy access

Demo

Here's the link to my ai agent:

Surfer 2: The Next Generation of Cross-Platform Computer-Use Agents - H Company

H Company builds models, agents, and products that automate tasks and simplify complex work. We empower people and enterprises to move faster, think bigger, and do more of what matters.

hcompany.ai

Flashcards created by StudySherpa:

StudySherpa – Flashcards – Photosynthesis - Google Docs

docs.google.com

StudySherpa – Flashcards – World War I - Google Docs

docs.google.com

How I Used Runner H

Tools Connected:

Google Docs (read/write)
Google Calendar (write)

Prompt Setup

Below is the full prompt used in Runner H for StudySherpa:

You are StudySherpa, a student’s personal academic assistant. Follow these steps in order:

1. Read all uploaded Google Docs. Assume they contain either lecture notes or a syllabus.
2. For each doc, extract important facts, concepts, or questions. Convert these into spaced repetition-style Q&A flashcards.
3. Create a new Google Doc for each source doc titled: “StudySherpa – Flashcards – [Original Title]” and write the flashcards in a clear Q&A format.
4. If the original doc includes any dates or deadlines, extract them and create Google Calendar events with those dates and a relevant title.
5. At the end, create a summary Google Doc titled "StudySherpa – Flashcard Index" and list all the flashcard doc titles and links in a bullet-point format.

Formatting for flashcards:
- Use headings for topic sections
- Bold the question and italicize the answer
- Example:
  **What is the powerhouse of the cell?**
  *The mitochondrion.*

Make sure everything is clear, organized, and actionable for the student.

How It Works

Step-by-Step Workflow:

Upload lecture notes or syllabus as Google Docs
Runner H reads them, extracts deadlines and knowledge
Flashcards are created in new Google Docs
Events are created in the user’s Google Calendar for key dates
Index Doc is created linking all flashcard documents

Diagram:

Student Docs ➔ StudySherpa (Runner H)
           └──> Flashcard Docs (Google Docs)
           └──> Calendar Events (Google Calendar)
           └──> Flashcard Index Doc

Use Case & Impact

This is perfect for:

Students with messy folders full of study docs
People studying for exams who want to review with flashcards
Anyone who receives syllabi with unclear deadlines

Impact:

Saves hours of manual note summarization
Keeps everything in one place
Reduces anxiety about forgetting tasks or dates

Try It Yourself

You can recreate this by:

Connecting your Google Docs and Google Calendar to Runner H
Copying the prompt above into a new Runner H agent
Uploading your study notes or syllabus docs manually

Social Love

// Detect dark theme var iframe = document.getElementById('tweet-1938517937377435984-605'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1938517937377435984&theme=dark" }

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

linkedin.com

Conclusion

StudySherpa V4 is just the beginning. In future versions, I'd love to:

Export flashcards to Anki
Add mind maps or concept diagrams
Chat with a GPT-based tutor based on your notes
Form collaborative study groups with shared progress

Thanks to the Runner H team and DEV for this challenge. This was a fun, practical way to explore AI agents for real-world student productivity ✨

📈 AdWiz: AI-Powered Ad Ideas, Straight to Slack

Sai Shravan Vadla — Wed, 25 Jun 2025 08:21:16 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

What I Built

AdWiz is an AI agent built with Runner H that automates ad ideation for creative and marketing teams. It scrapes trending ads and marketing insights from the web, compiles them into a structured Google Sheet, and sends a curated drop of ad ideas to your team’s Slack workspace—every week or on demand.

It solves the problem of creative block and research overhead for marketers by delivering relevant, timely inspiration right where teams already collaborate.

Demo

Here's the link to my Runner H run:

Surfer 2: The Next Generation of Cross-Platform Computer-Use Agents - H Company

H Company builds models, agents, and products that automate tasks and simplify complex work. We empower people and enterprises to move faster, think bigger, and do more of what matters.

hcompany.ai

And the spreadsheet it created:

Trending Ad Ideas - Google Sheets

docs.google.com

How I Used Runner H

Runner H made it incredibly easy to orchestrate a workflow across multiple tools with no backend or server to maintain.

Here’s how I built it:

Trigger: Manual run via Runner H dashboard.
Agent Prompting: The agent analyses current trends in advertising and brainstorm creative ad concepts.
Google Sheets Integration: The agent writes the ideas into a Google Sheet, formatting them into columns like Trend and Description.
Slack Integration: Once the sheet is ready, a message is sent to a specific Slack channel with the link and a quick summary.

Replicate it in Runner H:

Add Google Sheets and Slack as integrations.
Use this prompt "Create a Google Sheet with trending ad ideas and send it to my team on Slack".

Use Case & Impact

This is especially useful for:

Marketing teams needing fresh ad ideas regularly.
Creative agencies looking to stay ahead of the curve.
Solo founders and indie hackers who want to speed up content ideation.

Instead of spending hours browsing YouTube ads, X/Twitter threads, or marketing newsletters, teams can get a curated list of high-potential ad concepts automatically delivered to Slack—saving time and sparking creativity.
It transforms a chaotic, manual process into a reliable, repeatable workflow with just a few click

Social Love

// Detect dark theme var iframe = document.getElementById('tweet-1937788268541862285-375'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1937788268541862285&theme=dark" }

View this post on Instagram

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

linkedin.com

Meeting to Google Docs

Sai Shravan Vadla — Fri, 20 Jun 2025 16:53:33 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

What I Built

I built a super simple but surprisingly useful Runner H agent called “Meeting to Google Docs”. The goal? Just type in the details of your meeting, and boom — Runner H creates a clean, ready-to-use Google Doc titled with the meeting name and structured with your agenda.

This agent is for anyone (like me!) who constantly has to jot down meeting notes, agendas, or brainstorms and just wants to skip the repetitive formatting and file creation. It takes 10 seconds instead of 10 minutes.

Demo

Surfer 2: The Next Generation of Cross-Platform Computer-Use Agents - H Company

H Company builds models, agents, and products that automate tasks and simplify complex work. We empower people and enterprises to move faster, think bigger, and do more of what matters.

hcompany.ai

How I Used Runner H

Here’s how I set it up:
Connected my Google Docs account to Runner H.
Prompted Runner H like this:
“Create a Google Doc titled ‘[Meeting Name]’ with this agenda:
Team updates
Project progress
Roadblocks
Next steps”

Runner H instantly created a well-formatted Google Doc in my Drive — no extra clicks, no templates, no chaos.

You can easily replicate this by typing your own meeting title and points — that’s it! No code, no learning curve.

Use Case & Impact

If you’ve ever scrambled 2 minutes before a call to create a doc, or juggled messy notes across apps, this is your fix. It’s great for:
Freelancers and teams documenting client calls
Students prepping group work agendas
Startups managing stand-ups or sprint planning
Anyone who wants to stop reinventing the wheel every time they need a document

It doesn’t try to do everything — just one thing really well. And that’s the magic. ✨

Social Love

I’ll be sharing this on Twitter and LinkedIn soon — will update with links here once it’s live!

🏠 Mailo: AI Email Replies, Seamlessly Delivered

Sai Shravan Vadla — Sat, 07 Jun 2025 10:05:52 +0000

This is a submission for the Postmark Challenge: Inbox Innovators.

What I Built

I built Mailo, an email-based AI assistant powered by Postmark and OpenAI’s GPT-4. You simply send an email to the designated mail address, and Mailo replies with a thoughtful, relevant, and human-like response within seconds. It feels like you're emailing a helpful friend who’s really good at everything. Whether you're asking a technical question, seeking advice, or just curious about something—Mailo reads your message, understands the context, and sends a high-quality AI-generated reply.

The magic is in the simplicity: you don’t need to sign up, install anything, or learn a new interface. Just send an email like you normally would. You can also reply to Mailo’s responses to continue the conversation—threaded replies are fully supported. It’s as natural and frictionless as having an ongoing chat, but over email. The goal was to make AI feel invisible and effortlessly integrated into a tool everyone already uses daily: email.

Demo

mailo-nine.vercel.app

The landing page:

Sending an email to Mailo:

Receiving an email from Mailo:

Code Repository

shravzzv / mailo

Mailo is an AI-powered email assistant that lets you ask questions and get intelligent replies — all through email. No app, no login, just email magic.

📬 Mailo

Mailo is a lightweight AI-powered email assistant that works entirely over email. No apps. No logins. Just send an email and get smart, context-aware replies delivered straight to your inbox — powered by GPT.

Built for the dev.to Postmark Challenge: Inbox Innovators.

✨ Features

📩 Email-first Interface — Just send an email to interact, no account needed.
🧠 GPT-Powered Intelligence — Smart, human-like responses tailored to your message.
⚡ Fast & Lightweight — No bloated UI, just good old email.
🛡️ Privacy-first — Only the message content is used to generate replies.

🛠️ Tech Stack

Tech	Role
Next.js 15	API routes + landing page
Postmark	Inbound email parsing
Nodemailer	Sending reply emails
OpenAI (via GitHub Models)	LLM backend
TypeScript	Type safety

📦 Architecture

User → (sends email) to Postmark → Next.js Webhook → GPT-4.1 (GitHub) → Mailo Nodemailer → replies

🧪 Testing the App

Click the Try…

View on GitHub

How I Built It

🔧 Tech Stack

Mailo is built on top of modern, minimal technologies designed for speed, privacy, and simplicity:

Frontend: Next.js (App Router) + Tailwind CSS for styling and responsive UI.
Backend: Next.js API routes for handling outbound mail, webhook payloads, and AI response generation.
AI Engine: OpenAI’s GPT-4 for smart, high-quality email replies.
Email Infrastructure:
- Postmark's inbound stream for receiving emails.
- Custom MX record setup to route all emails sent to contact@shravzzv.tech directly into a Postmark Inbound Stream, which then hits a webhook in my app.

🧪 Testing instructions:

Testing Mailo is easy. All you need is an email account:

Go to mailo-nine.vercel.app and click Try it now, or just send an email directly to contact@shravzzv.tech.
Use a clear subject line as your question.
Optionally, add extra context or details in the body of the email.
Wait a few seconds — Mailo will respond to you automatically.
Got follow-up questions? Just reply to the email you received — Mailo understands and continues the conversation intelligently.

⚠️ Note: If you don’t see a reply, check your spam folder. All replies are sent within a few seconds.

♿ Accessibility

I made a conscious effort to ensure Mailo is as accessible as possible:

✅ Semantic HTML and ARIA-aware elements
✅ Keyboard navigability across all interactive components
✅ Responsive and legible on mobile, tablet, and desktop
✅ Email-first design — users who don't visit the website still get the full Mailo experience

✨ Highlights & Learnings

I learned to configure custom MX records and route emails directly through Postmark's infrastructure — giving me complete control over inbound email flows.
I built a robust, privacy-first workflow that requires zero user data storage.
I realized the power of using email as an API — the interface is already everywhere, no onboarding needed. -It was incredibly fun (and challenging) to make email feel conversational and alive with AI.

Thanks for checking out Mailo — I'm excited to see where this goes next! 👋 If you have feedback, questions, or ideas, feel free to… well, just send me an email. Built for the Postmark Inbox Innovators Hackathon with love.💛

The Social Roastmaster

Sai Shravan Vadla — Mon, 27 Jan 2025 04:06:42 +0000

This is a submission for the Agent.ai Challenge: Assembly of Agents (See Details)

What I Built

Social media profiles are a goldmine of comedic content, just waiting to be unleashed. But crafting the perfect roast, one that's funny and personalized, can be tricky. That's where The Social Roastmaster comes in. This AI agent analyzes your Twitter, Instagram, and LinkedIn profiles, identifying the funniest aspects of your online persona. It then uses that information to generate side-splitting roasts that are guaranteed to tickle your friends' funny bones. It also generates a meme based on your roast. Enjoy.

Demo

The Social Roastmaster

This ingenious agent leverages the power of four other AI agents:

Agent.ai Experience

Building The Social Roastmaster with Agent.ai was a delightful experience. The Builder platform was intuitive and user-friendly, allowing me to focus on the creative aspects of the agent.

The Social Roastmaster is a testament to the power and potential of AI-driven humor. By harnessing the capabilities of Agent.ai's platform and various AI agents, The Social Roastmaster injects a dose of laughter and lighthearted competition into your social media interactions.