Most people go to SF for the weekend and hit Dolores Park.
We went and built an AI that might save you $15,000.
Four builders. Six hours. One hackathon.
Here’s the backstory.
The problem nobody told me about
I’m thinking about moving to SF. So like any normal person, I started apartment hunting.
Found a place. $3,150/mo. Beautiful photos. Great vibes. I’m in.
Then I looked closer.
- $275 — parking (nearest garage)
- $140 — utilities
- $35 — trash fee (buried in clause 14.3, lol)
- $45 — pest control (mandatory)
- $15 — package locker (can’t opt out)
Real rent: $3,660.
That’s $6,120 a year nobody told me about. A round trip to Tokyo I’ll never take.
It gets worse.
The gym I picked the whole neighborhood for? 25 minutes by transit. I’d quit by month two and we both know it.
Management company runs 14 buildings. Tenants in 9 of them say repairs take 3 weeks.
The block gets loud Thursday through Saturday.
I’m a light sleeper.
Here’s the wild part: all this info is public. Building violations. Noise complaints. Tenant reviews on Reddit and Google.
It’s just scattered across ten sources nobody has time to check.
So you sign the lease. Cross your fingers. Find out the hard way.
The hard way costs $15,050. (Early termination + forfeited deposit + new deposit + movers + overlap rent. Fun.)
What we built
So this weekend, four of us said: screw it. Let’s build the thing that should already exist.
We built Nyx. Know the city before it knows you.
Paste a listing. 60 seconds. Nyx pulls:
- Building violations
- Block-level noise data
- Management reviews across their entire portfolio
- Hidden fees buried in 40-page leases
- Parking, transit time, amenities matched to your life
No dashboard. No scores. It just talks to you like a friend who’s lived in SF for ten years and won’t let you sign a bad lease.
It walks you through your actual Wednesday — gym, commute, coffee, when the block gets loud, what dinner looks like.
A movie of your life there. Not a spreadsheet.
Then one call:
- Go — move fast.
- Negotiate — here’s your leverage.
- Walk away — here’s why.
People make $40K decisions on listing photos and vibes.
We built the antidote to vibes. In six hours. On a Saturday.
What I actually learned
The idea that wins a hackathon is never the most technically impressive one.
It’s the one that makes a judge go: “wait, why doesn’t this exist already?”
That’s the whole game.
Most teams pitch features. Wedge angles. Smart architecture. Cool models.
The audience didn’t love Nyx because it was clever.
They loved it because most of them had personally been burned by exactly this problem.
That’s the lesson.
Pitch a problem the room is feeling. Build the thing that should already exist.
The other lesson
You don’t win a hackathon with a good idea.
You win it with people who can build under pressure without ego getting in the way.
Sreenidhi, Qimei, Makoto — every one of them shipped. No “let me think about it.” No “but actually.” Just: pick the cut, run the cut, ship the cut.
Six hours is not enough time for opinions.
It’s enough time for one team and one direction.
One more thing
We posted the idea on LinkedIn before the hackathon even ended.
People didn’t just react. They sent DMs. “I literally just got burned by this.” “My last apartment did this exact fee thing.”
That’s what gave us the confidence to pitch like we meant it.
When the room is already nodding before you finish your first sentence, you don’t need a clever pitch. You just have to not get in the way.
Try it
The prototype is live: vibe-perspective-spark.lovable.app
Paste any SF listing. See what it tells you.
If you’ve ever moved somewhere and thought “why did nobody tell me” — that’s who we built it for.