Launching today
Trag Studio
Local AI studio for unlimited images, video and music
10 followers
Local AI studio for unlimited images, video and music
10 followers
TRAG STUDIO turns your Windows PC into a simple local AI studio for images, video, and music. Generate unlimited images for free, without cloud credits, subscriptions, or node graphs. Upgrade once to unlock local video and music generation. Your files stay on your machine, workflows run in one clean app, and the experience scales with your own hardware.
15 years in, and this week I still felt like a first-timer.
I've shipped VR and real-time 3D products before — I know what it is to fight a frame budget, to make a headset render something your body believes. But generative AI was a different battlefield, and this was my first real fight on it. Trag Studio began as a side quest — the dream you open at 1am after the main studio's work is done. A generative studio that runs 100% on your own GPU. No cloud, no tokens, no $50-a-month-per-tool tax. Pay once, own it.
But the thing I care about most isn't the engine — it's the door. There's a whole crowd of people who feel the walls of ChatGPT and Gemini — the limits, the filters, the "I can't do that" — and want out… then take one look at the node graphs and setup rituals of today's "pro" AI tools and quietly give up. Trag Studio is built for them: a genuinely accessible interface sitting on top of a serious local engine. You shouldn't need to be an engineer to make images, video and music on your own machine.
Here's the humbling part: the models were never the hard part. The war was the last mile — the craft of making power feel simple.
Our installer would sometimes leave PyTorch amputated of its 2.5 GB of CUDA DLLs, so the GPU quietly vanished and the app fell back to CPU — for a while our own software was telling paying users their machine wasn't good enough, when the bug was ours. An expired auth token silently ate a customer's Pro unlock after he'd downloaded 100+ GB and paid. A tester literally called me — "will your fix let me use video?" — and that one call reshaped a release. On launch day, our payment provider didn't clear in time, so we migrated the whole checkout in a few hours, live, hearts in our throats.
And the part I'm quietly proud of: we built an MCP layer — so you can point an AI agent, yes, Claude itself, at Trag and let it generate media for you, locally. The assistant you already talk to, finally able to make images and video on your GPU.
The scariest part was never the code. It's the quiet fear that the idea won't be seen for what it is. Everyone says "execution is what matters, not the idea." So we executed. We shipped V1.0. One hour in — still shaking a little — I'm equal parts terrified and proud.
If you believe creators should own their tools instead of renting them: try it, break it, and tell me what Trag should become next. This time, I'm actually listening.
— Mickelly, CTO
@mickelly_edna The positioning was easy to explain. Making all that local AI power actually feel simple was the real battle.
Was there in the 1 a.m. builds, the impossible bugs, the launch-day payment migration, and the moments where it nearly broke us.
Proud to see your side quest become a real product people can finally use.
Now we listen, fix, and keep shipping.
Hey Product Hunt — Allan here, co-founder and CMO of Trag Studio.
We built Trag Studio because creative AI currently gives people two frustrating choices.
You can use convenient cloud tools, but every experiment consumes more credits, tokens, or another subscription. The more you explore, the faster the bill grows.
Or you can run AI locally — and suddenly find yourself learning models, connecting workflows, and diving in an entirely new technical discipline before creating anything satisfying.
The cloud keeps the meter running. Local AI often makes you become the engineer.
We wanted a third path.
TRAG turns your own PC into a simple local studio for AI images, video, and music:
— Unlimited local image generation, free
— Video and music with one lifetime purchase
— No recurring generation credits
— No node graphs to build or maintain
— Your work stays on your own machine
We are a team of three from Mauritius, and we built this first public version in roughly two months.
Not because the product is finished (it definitely isn’t) but because we believed this idea deserved to exist now, not after another year of polishing in private.
Our next major steps are improving TRAG through real user feedback, building an integrated editor, and bringing the studio to Mac, targeting fall 2026.
We are also exploring how TRAG could take advantage of a new generation of compact AI systems, including AMD Ryzen AI Halo and NVIDIA DGX Spark, while staying true to the same goal: making local creation accessible without turning creators into infrastructure engineers.
Today, we need honest feedback more than polite encouragement.
Thanks for giving our small team a look.
Allan
Hey Product hunt - Sarobidy , CFO of TRAG STUDIO
This launch means much more than releasing a product. It's the result of an incredible bootstrap journey where every decision mattered.
Building a company in a new country for the first time pushed me far outside my comfort zone. At first, the unknown felt like a wall of ice. But once you choose resilience over fear, you realize that nothing is impossible—and your learning curve accelerates faster than you ever imagined.
Those challenges shaped our product. We built it to solve the frustrations we experienced ourselves: token limits, usage restrictions, and concerns about data privacy. We wanted to create an AI that's simple, accessible, and gives users back control over their data.
Seeing this vision come to life makes every challenge worth it.
@sarobidy_edna You did amazing, handling all of the trouble during this incredibly tought journey