At Upcar, engineer caught an asymmetric-drift bug in tripPriceHelpers.ts — a host was overcharged $493 because the rounding step ran on the wrong side of the split. Drafted the fix as a PR. Founder reviewed, shipped to development, refunded the host — same day it happened.
Zavi runs Upcar. Upcar built Zavi.
Upcar is a P2P car-rental marketplace — a Turo competitor, live in 5 states, ~$100K monthly GMV, ~600 hosts. The agent fleet on this site is not theory. Every agent shipped because Upcar needed it that week. The architecture generalizes — the wedge was specific.
The agents weren't a product. They were an emergency.
Solo founder. ~600 hosts. Five states. Chargebacks at 11pm. KYC escalations from three time zones. Stripe drift on Friday afternoons. Each agent was a thing the founder built because he couldn't be in two places at once.
Then the founder noticed the agents were a moat — not the marketplace, not the brand, not the hosts. Ten-plus months of operating context — every state launch since July 2025 — informed how each agent was designed: the shapes, the envelopes, the decision classes, the failure modes the runtime had to handle. Zavi is what happens when you turn that context into a product.
| Jul 2025 | Upcar launches in CA. P2P car rental — a Turo competitor. |
| H2 2025 | Expands to NV, TX, AZ, FL. Five-state ops. Founder hand-writes the dispute-response runbook and per-state KYC rubrics through the first chargeback waves. |
| Q1 2026 | Booking-risk screen wired: Stripe Radar + Persona + account age. Stops a $4,200 Audi RS5 booking from a 3-day-old account. Founder realizes the ops playbook is the moat — not the marketplace. |
| Apr 2026 | Anthropic ships SKILL.md. Tom Blomfield writes the YC Company-Brain RFS. The shape of the moat snaps into focus. |
| May 2026 | Zavi spins out. The agents become the starter fleet. Onboarding the first operators. |
The kind of thing that happens on a real operating company.
Sanitized but real. Trip numbers and dollar amounts come from the actual Upcar history.
dispute-response shipped Stripe CE 3.0 packages on three trips between 11pm and 2am — pulling GPS, check-in photos, host messages. Founder reviewed in the morning. Two won, one lost.
support-lead flagged 6 accounts targeting Audi/Porsche listings, all from one device fingerprint. Proposed a Sift policy diff. engineer shipped it the same week. Zero successful bookings from that cluster since.
qa ran the suite on an OAuth-callback change inside the sandbox and surfaced a failing security assertion human review had missed — with the suspected root cause. Queued the bug-file decision. Founder rolled back, fixed, re-shipped.
Your company will produce the same kinds of decisions. Zavi turns them into agents.
Onboard your company →