geethikaisuru

Built Quickthrottle & Mahasen end to end, marketed em, got customers, all while in College.

In August, we got to the Top 20 teams at a Global UN Hackathon, building an AI for Digital Trade Regulatory Analysis. I did the engineering from architecture(others helped) to coding up and deploying. 2 other team members are Professors, 2 lecturers, and I was the only undergrad.

https://geethikaisuru.com/portfolio

- G.

i launched https://Quickthrottle.app on Product Hunt 2 days ago. No hunter, no launch list, no upvote groups. Just me posting it and sharing the link.

I tracked all the traffic with PostHog analytics. So I can see where each person came from and where they dropped off. Here's what the first 24 hours looked like 10 upvotes around 20 people came to the site from Product Hunt 2 of them clicked checkout 1 bought it

More info in the link..

I'm graduating in 23 days. I built few things while in college. But only got to $283 in rev. So Now I've 10 days left in September to reach $2000 in revenue or die trying.

I built Quickthrottle.app - A local content factory. You can build hundreds of reels & slideshows in minutes, driven by Claude code or codex. It's local & fully customizable. This is way better than those advertising on Reddit.

What you can do:

- Download & Remix Viral slideshows in 3 clicks - Make thousands of reels in minutes. For $0, forever - Use your Claude or Codex subscription to drive it - Lifetime ownership, no subscription - Fully Local & Customizable - One-time frontend

I was using this as an internal tool so far & decided to package it up as a one time product. This instagram account got 5M+ views using it: https://quickthrottle.app/blog/5m-insta/

It's a content generation & a multiplying tool. Not a discovery tool.

I had a goal of reaching $10,000 in revenue before the graduation even before I came to college. But my failure to reach it so far leave me with no choice. I know this is because I'm in Sri Lanka, & the environment is fully different. But excuses are worthless. I've not been sleeping properly for the entire month. I won't be sleeping for the rest until I hit the goal of $2k rev before Sep 30th. If I can't, I will sleep forever on October 1st.

God bless.

I'm graduating in 24 days. I built few things while in college. But only got to $283 in rev.

So Now I've 10 days left in September to reach $2000 in revenue or die trying.

I built Quickthrottle.app - A local content factory. You can build hundreds of reels & slideshows in minutes, driven by Claude code or codex. It's local & fully customizable. This is way better than those advertising on Reddit.

What you can do:

- Download & Remix Viral slideshows in 3 clicks - Make thousands of reels in minutes. For $0, forever - Use your Claude or Codex subscription to drive it - Lifetime ownership, no subscription - Fully Local & Customizable - One-time frontend

I was using this as an internal tool so far & decided to package it up as a one time product. This instagram account got 5M+ views using it: https://quickthrottle.app/blog/5m-insta/

It's a content generation & a multiplying tool. Not a discovery tool.

I had a goal of reaching $10,000 in revenue before the graduation even before I came to college. But my failure to reach it so far leave me with no choice. I know this is because I'm in Sri Lanka, & the environment is fully different. But excuses are worthless. I've not been sleeping properly for the entire month. I won't be sleeping for the rest until I hit the goal of $2k rev before Sep 30th. If I can't, I will sleep forever on October 1st.

God bless.

So, I shot one UGC clip and it ended up doing around 5 million views across the posts I made from it.

To be clear:

- A text hook is the line of text sitting on top of the video. The text overlay. - A clip is the video underneath it. Usually a silent UGC style shot, and it must look natural. - One clip plus however many hooks you wrote is that many posts. Same footage every time.

The short version Map your audience. Who are they, what do they actually complain about. Find the things they care about. I scroll Reddit for this, or just TikTok comments. Feed those to Claude and get hooks out of it. Shoot one background UGC clip that fits. Drop the clip and the hooks into the Quick Throttle hook generator (free, runs in your browser) and it spits out a video per hook. Post. Now the details.

1.Mapping the audience Pick who your audience is. It's best if it's something you know about deeply. Don't ignore or skip this. Don't just start posting randomly. Think first. Use your brain. Don't ask AI to do this.

Reddit is the best place for this. Find the subs your people actually sit in and read the comments, not the posts. The posts are performative, the comments are honest. Save the lines verbatim. Don't clean them up. The exact wording is the whole point, that's the stuff that makes someone stop.

TikTok comment sections work too. Same idea. Just pick something you relate to and you know.

2.Find and multiply hooks I tried using my brain and coming up with hooks. AI is already shit at this. But the best results I got were when I used the same hooks that had already gone viral. There's this weird thing about hooks that went viral. They hit certain people every time. So find hooks that already went viral, preferably in the same format, single clip and different hooks.

Then use Claude to multiply those hooks. Use the Opus model at max thinking. First tell Claude what you are doing.

"I'm making this channel using this format for this audience. I guess these will be what I'll be talking about, and these are some hooks that we can use as the text overlays for the videos. And I'll be using an automation tool to make tens or hundreds of videos at scale. So I need you to multiply these hooks…"

Now you're 70% done.

3.Shooting the clip Don't overthink this, bro. Just shoot something. This will sit in the background. Just shoot it and get it over with. One clip. Silent. Something loosely relevant that can sit under any of the hooks without contradicting them.

Keep it vague enough to carry every line you wrote. If the clip is too specific it only works for two of them and you are back to shooting again.

A girl in the video works the best. I didn't see many with boys' reaction videos. Just shoot it. Don't overthink. Under 10 seconds.

4.Making the videos You can export the clip to CapCut and add the text one by one. Or just use the tool that I made. I used this exact tool to make the videos above, a few months ago.

Go to quickthrottle.app/hooks 1. Upload the clip 2. Copy and paste the hooks 3. Select the text style, and the position in the canvas 4. Render and download the videos

It all happens fully locally in your own browser. So no data is ever uploaded. It's part of a bigger content factory, and that's why it's completely free.

5.Posting If you're posting for the first time, don't post 4 posts on the first day. First warm up your account. Like, comment, follow and DM with people just like a normal guy for the first 4 days after making your account.

Then start posting 1 to 2 posts a day. Then gradually increase the frequency. But keep it to 3 to 5 posts per day at max.

Watch which hook goes, then write more hooks like that one and reuse the same clip again.

That's the actual loop. Your clip is the reusable asset, hooks are the cheap part. If something isn't working the answer is almost always more hooks, not more footage.

Hi team, I'm doing a user research on how early stage founders are doing marketing. Would be great to hear how you are doing it. Just a 10 minutes call where I listen to you on how you do it. Nothing else. Specifically looking for 1-3 person team founders who are actively building or starting to build something right now. Also I'll share everything I learn from interviewing everyone. Thank you so much & much appreciated team. <3

Looking for early adopters for a Proactive Marketing AI for indiehackers & solo founders. It's a voice dictation app which let's us do voice typing to Cursor/ Claude Code or any app. At the end of the day, the AI look at all the transcriptions, connect the dots and give you authentic storytelling posts based on your own experiences. Imagine you dictate a reply to an unique customer email, you spend 12 hrs building a feature. The app will pick it up and give you posts to share on social medias. No more AI slop because it's your stories. You automate only the storytelling. Just copy and share. | Join waitlist for early access:https://forms.gle/b32Rp9ctp6CL1FzF8

what's the alternative?

I hate when suddenly Claude code says, 5 hour limit reached. Would have been great to schedule my coding sessions. Got enough of waiting for Anthropic to build this. So I built a simple web app to track it. Enter your last known reset time and see 30 upcoming windows.

- Basic HTML and some style. - Fast & lightweight af - Data save in localStorage.

heh :)

Mahasen Voice is fast AI dictation & an assistant both in one. Just Hold 'Win+Ctrl' and speak anywhere on your PC. It adapts to each app naturally. AI mode can be started in any app by saying "Mahasen" to give direct instructions. If in email: "tell him I'll be 20min early."

Innovation should free us from complexity, not create it.

But for some reason, our lives are getting more & more complex. That's the exact reason behind this.

AI should handle the bullshit. I should have the time to do what truly matters. What makes us truly human. With Mahasen Voice, we're trying to take progress in that direction.

Mahasen Voice will truly help you. No bullshit. It have already saved me hours. I hope it'll do the same for you.