We've been building software professionally for over a decade, and we train. Yuge is what happened when we got tired of pretending the tools available were good enough.
The pattern that kept frustrating us was this: the fitness apps we used were built by people who either understood software or understood training, but rarely both. The software people built beautiful loggers that treated every exercise the same. The training people built spreadsheets with macros and called it an app.
Nobody was building an actual coach.
nahyeh is our attempt to fix that. We build AI products for specific industries, and we go deep. Products where the AI is actually doing the work that used to require a human expert sitting across from you. Yuge is the first product. Hoist — the agent layer that handles a gym's admin so coaches can coach — is the second.
The name comes from Australian slang. "Nah yeah" means yes. It's how Australians agree with something while pretending they're not that enthusiastic about it. The name felt right for a company that takes real problems seriously without taking itself too seriously.
The first product: Yuge
Yuge is an AI strength training coach. Not a logger. Not a template generator. A coach.
That distinction matters. The market is full of apps that store what you did and maybe add 2.5 kg when you hit your rep target. They don't understand training. They don't know the difference between a deload and a missed session. They can't distinguish fatigue from a real plateau. They don't know what RPE 8 means in the context of week three of a mesocycle.
We asked a different question: if you stripped away the constraints of time, client load, and memory, what does good coaching actually require? Then we worked backwards from that.
The answer is an AI agent, not a chatbot. Something with memory, tools, and the ability to make real programming decisions. It looks at your last six weeks, notices your deadlift volume has been creeping up while your logged RPEs are climbing faster than the weights, and pulls back before you dig yourself a hole. It understands 5/3/1, GZCL, DUP, block periodization, linear progression. It can explain every decision it makes.
Who it's for
Yuge is built for lifters who already understand training. People who've read enough about programming to be dangerous, who track their training, who get frustrated when an app tries to explain what a rep is. These lifters don't want a template. They want something that handles the programming so they can focus on the actual work. The AI doesn't replace their knowledge. It extends it.
Someone who's never touched a barbell can describe what they want and walk away with a real program built on legitimate periodization principles. And someone who's been training for five years gets something that meets them where they are, not a template designed for the lowest common denominator. The same tool serves both because the intelligence sits underneath, not in a fixed template on top.
Who we're starting with
We're launching with a small group of serious lifters first. The fitness culture we're building for is serious without being performative — gym culture that rewards results over aesthetics. That aligns with what we're building.
Being close to early users matters when you're still figuring things out. The lifters who get in early will have outsized influence over what Yuge becomes. That's not a marketing position. It's a practical one. Generic feedback produces generic software. Specific feedback from people who know what they're talking about produces something worth using.
Once it works well with this first cohort, we'll take it further. Not before.
What comes next
We'll write about what's working and what isn't as we go. No launch hype. If you want specifics on how the coaching intelligence works under the hood, start with how Yuge handles training interventions.
More soon.
