Built so my mom would actually use it.

SokoJin started with a broken dryer, a $340 service quote, and a YouTube video that didn't match the model in our garage. It turned out to be a $14 thermal fuse. Fifteen minutes of work, once I knew. That gap, between "I have no idea what's wrong" and "oh, that's all it was," is the gap SokoJin is trying to close.

Why "Jin"?

Because "AI assistant" is dead on arrival. Jin's a character: patient, a little dry, will admit when they're not sure. The kind of friend you'd actually want next to you when you're elbow-deep in a dishwasher.

Who's building it

I'm Henry Moseti (Founder). The guy in the family group chat who gets every photo of a broken thing with three question marks. Not because I'm trained. Because I'm stubborn enough to take things apart, patient enough to figure out what went wrong, and I've accidentally saved a lot of friends a lot of money over the years. SokoJin is me trying to make that less accidental, for anyone who doesn't happen to have a Henry in the group chat.

I'm building it in public because the only way an AI repair assistant gets good is by being wrong in front of real people and learning. Every thumbs-up/thumbs-down trains the next version.

Why a website with affiliate links?

Two reasons, and I'll be direct about both:

1. The guides are the product preview. If they help you, you'll trust Jin more in the app. So I write the best repair guides I can, for free, with parts lists.
2. Amazon affiliate revenue funds the build. When a guide links to a $15 thermal fuse and you buy it, I earn a few cents. That pays for the API bills that keep Jin running while we're still pre-revenue. Full disclosure here →.

What's next

First cohort of testers (TestFlight) opens once the waitlist hits 100. After that: more guides, more device categories, and — if there's signal for it — a paid tier for people who want priority access to a human pro when Jin can't help. Join the waitlist and tell me what's broken; that's the most useful thing you can do today.