Case Study
Chad Blumenthal
'25
5 min
TL;DR
I'm an entrepreneur at heart. Back in 2019, I left my job in investment banking to start my own company. It was a social meetup app called Tribe. Good idea, but terrible timing. 3 months after launch, COVID hit and shut down all social life and in-person contact around the globe…not good. While Tribe failed, my creative itch and builder mentality never went away.
That's what drew me to the field of product management. And over the past year, AI has redefined what's possible for PMs. I wanted to put the latest AI tools to the test and see how far I could make it, building solo. That's where VibeTools was born.
VibeTools is a lightweight, zero-cost alternative to Intercom that integrates directly with Slack. It's built for vibe coders, by a vibe coder. With a one-shot prompt, you can install a support widget to your product and in minutes, start talking to customers without ever leaving Slack. Oh, and it is beautiful and customizable with gradients and glassmorphism themes to match any brand.
I'm amazed at what I was able to build solo with Cursor in just a few weekends. It's a fully functioning app, hosted on Vercel, with a working backend on Supabase, and integrations with Slack. In this post, I’ll cover the journey of building VibeTools, including:
The problem
The solution
How I built it
Key challenges I faced
What’s next
Reflections
👉 Want to check out VibeTools? Try it now at vibetools.fyi
The Problem

Vibe coding is exploding. Tools like Cursor and Supabase make it possible to ship a fully working product in days, not months. For the first time, anyone, regardless of technical background, can become a builder. However, most SaaS tools are not catered to serve solo builders and small teams that are launching projects in the AI era. We're seeing vibe coders build weekend projects and scale to millions of revenue in record time. They need tools that are cheap, easy to set up, and simple. Vibe coders want to move quickly, but SaaS is expensive and bloated with unnecessary features.
For vibe coders, shipping has become much easier. But launching your new product is not the end of the story. The most important thing isn’t the code, it’s talking to customers, iterating on feedback, and improving over time. And here’s the problem: existing support platforms like Intercom and Zendesk are expensive, bloated with enterprise features, and require technical resourcing to set up. They force you to work in external ticket systems and don’t fit the modern, AI workflow of vibe coders.
That's why I built VibeTools, a lightweight, zero-cost alternative to Intercom that integrates directly with Slack. Set up with a single one-shot prompt in your favorite AI code editor, and in minutes, start talking to customers without ever leaving Slack.
The Solution

Support should feel effortless…for customers, for support teams, and for builders. With VibeTools, conversations with customers don’t disappear into a ticketing system that has to be carefully watched over. They show up exactly where your team already works: Slack. For the first time, talking to customers feels as natural as messaging a teammate. And with a single prompt, builders can set up a beautiful support widget on their website in minutes.
Here’s how it works:
For Builders
Sign up for free at https://vibetools.fyi
Customize your support widget with beautiful gradient and glassmorphism defaults
Add the below one-shot prompt in your favorite AI code editor like Cursor
Install the VibeTools Slack bot to start receiving/sending messages
For Customers
Real-time customer support
Supports emoji + markdown
A modern, responsive chat widget
For Support Teams
Each customer gets their own dedicated Slack thread
Reply directly in Slack… no dashboards or support tickets
Automatically collect emails from website visitors
How I Built It

I didn’t start this project in Figma. There were no production-ready designs and no technical collaborators. I started by opening up Voice mode in ChatGPT and began talking. It was a 2-way brainstorm session, where I shared my experience with the problem, posed potential solutions, researched existing tools in the market, and stress tested ideas.
From this conversation, a first cut of a PRD was created. After a few iterations I jumped straight into building. I fed the PRD into Figma Make to create a quick prototype and sense check the UX. After a few adjustments, I was satisfied and moved to coding. I gave Cursor the updated PRD, key screenshots from the prototype, and asked it to build my app.
In a single weekend, I was able to go from vision to a live landing page and working front end. I was blown away - what usually takes weeks and requires designers and engineers collaborating I could do alone in a single weekend.
As any vibe coder will tell you (for now), getting the first 80% is easy with tools like Cursor. That last 20% is still hard and requires a lot of iteration and patience - more on that below.
I persisted through all of the blockers and hallucinations, with the occasional use of Max Mode and a frustrated tone when talking to Cursor to really break through. Within 3 weekends, I had a fully functioning product, integrated with Slack and working backend. I was stoked. If you're curious, here is the tech stack:
Frontend
Lightweight React widget SDK, embeddable via a simple script
Responsive CSS with gradients, glassmorphism, and solid themes out of the box
Real-time updates via polling (WebSockets coming soon)
Backend
Next.js 14 API routes
Supabase Postgres for the database
Slack Web API handling instant, bidirectional threads
Infrastructure
Vercel for hosting and deploys
Supabase for authentication + database
GitHub for version control and CI/CD
Other Tools
Figma Make for prototypes
Cursor for AI code editor
ChatGPT for brainstorming and PRDs
Slack for product testing
Key Challenges

1. The last 20% problem
With AI code editors like Cursor, non-technical users can get 80% of the way to a working product very quickly. But the last 20% is where things get hard. Things like setting up backend infrastructure, connecting external APIs, and debugging integrations is quite challenging. This part requires persistence and iteration. Over time, I expect AI tools will smooth out this complexity, but for now, grit and patience are essential.
2. Emoji mapping
Slack isn’t Slack without its emojis. I wanted users to be able to send an emoji in Slack and see it render properly in the support widget. Little did I realize Slack has thousands of custom emoji that need to be imported and transformed to Unicode emoji characters. To make this work, I integrated an open-source emoji dataset (github.com/iamcal/emoji-data) and mapped thousands of emoji into the widget. It was a surprisingly fun but intricate piece of the build.
3. Thread management
One of the core design decisions was to let teams handle support directly in Slack. Talking to customers is as easy as sending a Slack message. That meant each customer needed a persistent thread in the #support channel. To enable this, I stored thread IDs in Supabase and mapped them to customers so that every conversation maintained its full history and context.
4. Real-time synchronization
For support to feel natural, messages had to flow instantly between the widget and Slack. Slack webhooks alone weren’t reliable, so I implemented a hybrid system: instant API calls from widget → Slack, and webhook events plus a 10-second polling fallback for Slack → widget. All conversations are persisted in Supabase for reliability.
What’s Next
After a few weekends of building, the MVP is now complete. This was never about launching a new product, it was a test to see how far I can take modern AI code editors like Cursor without any technical help from engineers or designers. I was blown away with what I was able to accomplish and more energized than ever to be a PM.
As for next steps with VibeTools, I will ask a few friends building side projects to test it out and give me feedback. In my spare time, I might work on new features top of mind:
Integrations → Extend beyond Slack with Discord and SMS support
Collaboration → Mentions and internal notes
Customization → Deeper branding and styling options out of the box
Automation → Knowledge-base integrations and an AI-powered support agent
Reflections
For me, VibeTools wasn’t just a weekend project. It was proof of how product management is changing. I went from idea → PRD → production in a few weekends, without design resourcing or waiting on engineering cycles. The exciting part is what this means for PMs. With AI tools like Cursor, the gap between idea and working product has collapsed. That means it's faster to validate ideas, iterate with confidence, and deliver more impact with fewer resources. And that…is pretty cool.
👉 Want to check out VibeTools? Try it now at vibetools.fyi









