Excited to announce LeadJot ✨ Ever lose a lead because you didn't reply fast enough? LeadJot is an AI chatbot that responds to your leads instantly and books appointments automatically. Even at 2am. Even when you're in meetings. You can set it up in 10 mins. Build your FREE Chatbot -> leadjot.com

Many don't know but I have this local fuel price tracking website for Georgia :D petrol.com.ge

Excited to announce ThumbnailPilot 3.5 - Image Editor, New Design and more... ✨ This update brings a full thumbnail and title studio redesign, AI image editor, title session follow-ups, activity notifications, and core improvements. Try Updated ThumbnailPilot now → thumbnailpilot.com

How come i have 605 followers here and only 240 on X. Let's fix it guys :)

Here's how I create viral shorts for my product: [thread 🧵]

ThumbnailPilot is officially live on ProductHunt! The goal is to create the simplest and most powerful thumbnail tool. And we are just getting started! link below 👇

Excited to announce: ThumbnailPilot 2.0 ✨ Generate Viral thumbnails in Thumbnail Studio in Seconds! 1️⃣ Describe your thumbnail 2️⃣ Upload reference images OR sketch your thumbnail in the app 3️⃣ Click Generate and get a viral thumbnail in a few seconds 👉 Try it out here → thumbnailpilot.com

How I've managed to humanize AI for writing content. 🧵 11 things to add your prompt (+ full prompt at the end):

Our developer brains think of marketing as a non-productive activity. We need to condition our developer brain to see marketing as productive. Shipping code without users is like writing a book no one reads. Marketing feels unproductive, but it's what makes the work matter.

So apparently gpt-4o soon will be able to generate entire lego set instructions. Imagine dropping a photo of your LEGO pieces and getting step-by-step instructions of what you can make. @LEGO_Group

Underrated way to ship apps insanely fast: 1) Build a custom boilerplate (auth/billing/account) in your fav stack 2) Reuse it to build ideas fast and validate 3) Kill what doesn't work and double down on what works Speed is your moat.

Nice so OpenAI used ChatGPT to create it's 34-page masterclass on building AI agents.

My fav stack to build apps: Server: - Python (FastAPI) - Postgres - Redis (auth, cache, rate-limiting) - Deployed on DO/Hetzner Client: - App: Vite React - Landing: NextJS static - Deployed on Cloudflare pages for FREE

Sharing this spreadsheet of 394 directories & 130 proven marketing tactics to boost your startup's Domain Rating. Submit to 2-3/day and grow your DR. link below ⬇️

Build a b2c app they said, it's gonna be fun...

Ouch, but true...

If you use ChatGPT, Gemini, Claude or Grok for writing, you should append this to your prompt to make it sound human. DO's: - Use conversational language and tone - Use a natural, concise language and sound like a human. - Use storytelling elements

Imagine standing in virtual queues to use certain AI tools. < 2024: standing in queues at supermarket 2025+ : standing in queue to use manus ai agent

Every time I use the web today, I always imagine what happens behind the scenes after a certain action...

Finally got invite to Manus! What should i try first?

How come I have 595 followers here but only 169 on X.

TIL 50k+ ppl/week can't generate random strings/numbers. Why does this exist...

Why does everybody prefer @DrizzleORM over prisma?

Meet ThumbnailPilot — Thumbnail Preview, Collaboration and Inspiration Platform ✨ Preview thumbnails in real YouTube feed, compare designs, collaborate with your team and get AI feedback before publishing. Ready to create thumbnails that stand out and improve your CTR? Click the link below 👇

God bless enums...

Many don’t know, but your app can get FREE OpenAI credits through Microsoft for Startups Founders Hub: - up to $150K in Azure credits - $2.5K in OpenAI credits Best use of this is deploying o-series llm models and building AI agents with that.

Finally got image gen working 🔥

Quick tip to save $ and llm calls on repeated queries: Cache responses in redis or just use Cloudflare AI gateway and enjoy auto prompt caching

Published first YT short promoting my app, is this any good? I read the goal is 70/30 split

As much as we hate Adobe, ngl they cooked with this updated UI...

Does any body remember ComfyUI?

Turns out you can use <break time="Xs" /> to add a natural pause in Elevenlabs...

TIL you need to pay $5 to register as a chrome web store developer

Unpopular Opinion: Cookie auth > JWT token auth JWT not designed for this purpose and is not as secure (especially if you "vibe-coders" store them in local storage lol).

Another things "vibe-coders" will usually mess up is concurrent ops, resource management, etc. But that won't matter, as 90% of apps would have bigger problems (security, constant bugs) long before those issues arise.

Found this super useful blog about UI/UX best practices. Highlighted a lot of great points that I am implementing right away. Link below 👇

Many don't know but there's actually open-source Redis alternative — Dragonfly. And it's also a lot faster :) ✅ Redis API compatible ✅ 25x more QPS than Redis ✅ 12x faster snapshotting

Launching my next SaaS in next few days 👀

Been learning how to build web widgets lately. It turns out to be both a somewhat complex and interesting process. A lot of stuff to learn: - Iframes vs Web Components - Fighting the bundle size (React -> Preact) - Bundling into single js file (rollup, webpack) - etc.

Preact is truly amazing. Same React app. One is bundled as is. Another uses preact/compat to alias React imports and swaps in Preact under the hood. Bundle size (gzip): React: ~140KB Preact: ~8KB All of this by just changing a few lines and installing a couple of packages.

340k lines...

Been shipping this year …

What's your go to terminal? I personally use Warp. Why? Warp has a lot of cool stuff built in: → AI Command Search → Warp Drive: Save and reuse command workflows → Warpify: Share terminal sessions → Blocks View: Auto-group commands and outputs Get warp here: app.warp.dev/referral/WLD...

I deployed my first model on Replicate. All whisper models were super slow or not super accurate. That's why I rewrote one of them to use batched whisper. Results: ~1,5 minutes per 1hr of audio (not counting the cold start) check it out on Replicate 👇

Hetzner is raising prices 😢 No more 20 TB bandwidth #buildinpublic

Haha, Reddit got it right 😂

brb doing marketing :) #buildinpublic

Space creation also got an upgrade 🚀 More ready prompts coming soon...