I built Tiramisu AI because I was tired of fighting with AI tools

I didn’t start Tiramisu AI with a big vision or a pitch deck.
It started with frustration.

Like many indie makers, I use AI a lot — images, videos, experiments for landing pages, product ideas, social posts. And every time, the experience felt… broken.

One tool for text-to-image.
Another tool for image-to-image.
A different one for video.
Different UIs, different pricing models, different limitations.

I wasn’t trying to build a studio-quality film.
I just wanted to turn ideas into visuals quickly.

But instead of creating, I spent most of my time:

• Switching between tools

• Rewriting prompts because each platform behaved differently

• Paying for features I barely used

• Hitting arbitrary limits that broke my flow

That’s when I realized the real pain point wasn’t “AI quality”.

It was friction.

The real problem I noticed

Most AI generation tools are built either:

• for researchers

• or for power users

• or as feature checklists chasing trends

Very few are built for creators who just want things to work.

I wanted something:

• Simple enough to start in seconds

• Flexible enough to go from text → image → video without starting over

• Powerful, but not intimidating

Something that feels less like “using AI”
and more like continuing your creative thought.

So I decided to build it

At first, Tiramisu AI was just for me.

I stitched together models.
Built a minimal interface.
Removed anything that felt unnecessary.

My rule was simple:

If this step breaks my creative flow, it doesn’t belong.

No fancy branding.
No growth hacks.
Just focus on making the experience smoother.

Over time, friends started using it.
Then strangers.
And their feedback was surprisingly consistent:

“It feels calm.”
“It’s the first time AI didn’t overwhelm me.”
“I actually finished what I started.”

That’s when I knew this wasn’t just a personal tool anymore.

Why the name “Tiramisu”?

Tiramisu isn’t complicated food.
But when done right, it’s layered, balanced, and satisfying.

That’s what I want this product to feel like:

• Multiple capabilities

• Layered together

• Without overwhelming the user

Where I’m at now

Tiramisu AI lets you:

• Go from text to image

• Image to image

• Text to video

• Image to video

• Video to video

All in one place, with one mental model.

It’s not perfect.
There’s still a lot to improve.
But it’s honest, focused, and built with care.

Why I’m sharing this here

I’m posting on Indie Hackers because this is where people understand building slowly, imperfectly, and intentionally.

No VC story.
No overnight success.
Just a tool born from a real problem I personally had.

If you’re building something because you needed it yourself — I’d love to hear your story too.

Thanks for reading 🙏