Built a MACH/VIX scoring engine — looking for first 5-10 users to break it
Built a scoring engine that grades market conditions (crypto, weather, anything time-series) on two axes — speed of change (MACH) and volatility/risk (VIX) — and I'm looking for the first handful of people to actually try it and tell me if it's useful.
What makes it different from the usual "AI benchmark" pattern:
- The AI-provider leaderboard on the site is built from real production calls, not a one-time synthetic test — a provider only shows up once it's actually been called
- There's a /calibration endpoint that checks the engine's own past predictions against what price actually did afterward
- Thresholds can be personalized per-symbol from its own historical data instead of one fixed cutoff for everything
It's live at machvix.ai. I'm not trying to sell anything here — I'll give free pro-tier access to the first 5-10 people who actually use it and give me honest feedback, good or bad. Comment or DM if you want in.
Scoring across domains is powerful. Curious — who feels the pain enough to rely on this daily?
Honestly? Nobody yet, and I want to be upfront about that rather than oversell it. Right now it’s a solo project with zero paying users — the honest answer is I’m still looking for the first person who feels that pain daily, not claiming I’ve found them.
What I have tried to build toward that: instead of a fixed “high risk” cutoff for every symbol, thresholds are computed from each symbol’s own history — so a normally-calm asset getting volatile actually stands out, instead of getting lost next to something that’s always wild. That’s a step toward “actionable” rather than “another number,” but I don’t have proof it works yet. That’s exactly what I’m trying to learn from people who’d actually use it.
Really respect the honesty here — that alone puts you ahead of most.
The dynamic threshold idea is actually interesting because it shifts from ‘data’ to ‘context,’ which is where most tools fall short.
The key now is finding someone who already makes decisions off this kind of signal.
In our experience, it’s less about proving it broadly and more about finding 1–2 users where it clicks hard.
Happy to help think through where to find those if useful.
The interesting challenge with MACH/VIX is turning a complex market signal into something that creates actionable decisions for users.
How are you thinking about proving that the scoring engine helps users make better decisions rather than becoming another indicator they monitor without acting on?
Fair challenge, and I don’t have a clean answer yet — which is probably the most honest thing I can say. Two things I’m leaning on to try to get there:
1. The /calibration endpoint checks the engine’s own past scores against what actually happened afterward. If a “high VIX” reading didn’t precede real volatility, that’s a failure I can see and fix, not hide.
2. Per-symbol thresholds (vs. one fixed number for everything) are meant to cut down on noise — fewer false “high risk” flags for things that are just normally volatile.
But neither of those proves someone changed a real decision because of it. That’s the gap. If you’ve seen tools cross that line from “indicator people watch” to “thing people act on,” I’d genuinely like to know what made the difference for those.
Appreciate the honesty and context.
Would be good to continue the conversation around how you're thinking about that gap as you validate the product.
What's the best email to reach you on?
semcry.intro (at) gmail (dot) com — also on the site's Contact section if that's easier. Looking forward to it.
Thanks! I’ve just sent it over.
Looking forward to hearing your thoughts whenever you have a chance.
Hey Aryan, thanks for reaching out — looks like the email hasn’t come through yet on my end. Could you double check it went to semcry.intro@gmail.com? Might be worth checking your sent folder too, just in case. Happy to take a look as soon as it lands!”
Thanks for the heads-up! I've just resent it. If you don't see it this time, would you mind checking your Spam or Promotions folder as well, just in case? Looking forward to hearing your thoughts.
Just replied to your email — might be worth checking your inbox (or spam/promotions, just in case) rather than waiting here.
Also, for anyone else reading this thread: not looking to purchase services here — open to investment or referral/commission-based partnerships, but that’s a different conversation than a paid engagement.