How it works

The whole product is built on one observation: your friends already know.

Most apps for finding people are built around reviews from strangers. Refind is built around something quieter and stronger — the fact that the same plumber's phone number keeps showing up in the contacts of the people you trust.

The whole product, in four moves

A trust graph builds itself, one quiet save at a time.

01
Save

You save the people you already use.

In Refind, you scroll through your contacts and tap the ones whose work you already trust — the plumber you call every spring, the dog walker, the moving company that didn't break anything.

Joe Mancini
Plumber
Lucia Park
Electric
Marcus Vu
Carpenter
Dani Wright
Roofer
02
Overlap

Refind quietly notices when two friends save the same phone number.

Saved phone numbers are normalized and matched by a one-way hash — never by readable number or name — and Refind surfaces the overlaps. When Maya and Sam both save (510) 555-0144, that number becomes a known worker in your network.

Maya Sam Joe Mancini (510) 555-0144 2 friends
03
Surface

When you need a roofer, your roofers are already ranked.

Type the trade, see the names. Each result shows exactly who in your circle vouched and how recently. No reviews, no star averages, no SEO. Just shared phone numbers.

roofer
Dani Wright
Maya, Aaron, Priya +1
4
Rose Cole
Sam, Priya
2
A&B Roofing
Aaron
1
04
Ask

You ask. The worker pays — not you — only if they accept.

A short note carries the names of the friends who vouched. If the worker accepts, they pay a small flat lead fee, and Refind keeps it.

Worker pays
Flat fee
per accepted job
Goes to
Refind
Flat fee, no commission
The unfair advantage

Reviews can be bought. Phone numbers in your friend's address book can't.

The reason fake reviews work is that strangers have no way to check. They read what's written and trust the average. Refind throws all of that out: the only signal it cares about is whether multiple specific people you know have a worker's number saved.

To fake that, someone would have to get into your friend's phone — twice, from different friends — and add the number. The cost of forgery is the cost of theft. The system can't be gamed.

×
Five fake Yelp reviews
Cost: $40. Time: an evening.
Two real people saving you
Cost: do two good jobs. Time: a few weeks.
Privacy, in plain English

Your contacts are encrypted, and only ever visible to you.

Refind matches people by a one-way hash of phone numbers — never by name. Your contacts themselves are stored encrypted and shown only to you.

What we store: your synced contacts — names and numbers — encrypted and paired with your account, visible only to you, plus a one-way hash of each number so we can spot overlap with other Refind members.

What we never do: sell or share your contacts, or show them to anyone else. There's no public directory, and matching is by hashed number, never by name.
On your phone
Maya's contact:
Joe Mancini · (510) 555-0144
Stored by Refind, encrypted
visible only to Maya · matched by a one-way hash:
8f4a · 1c9d · 6b2e · a73f · 5dd1 · 9c08 · 4b21 · e0fa
Overlap with other members
maya_uid → [
"8f4a1c9d6b2ea73f5dd19c084b21e0fa",
"a9027e4c9f2b…"
]
FAQ

Common questions

Does the worker know who vouched for them?
Yes — when a request comes in, the worker sees the names of the friends in your network who saved them. That context is what makes the lead feel personal instead of cold.
Do I need everyone in my contacts to be on Refind?
No. Refind only spots overlaps between people who are both on the app. Two friends is enough to start surfacing names. Each new friend who joins enriches your map a bit more.
Why hash numbers instead of matching on them directly?
Because a database of phone numbers is a juicy target. Refind builds the overlap graph on a one-way hash of each number rather than comparing readable numbers, and your contacts are stored encrypted at rest. We don't sell or share them, and they're never shown to other members.
How do you handle bad workers?
Customers don't rate workers publicly — there's no review system. The trust signal is whether their friends keep saving the worker. If a worker stops being saved, they quietly disappear from new searches in that circle.
What's stopping a worker from gaming this with friends?
A worker's own friends saving them looks the same as any other vouch — so initially they get some lift. But the system only really compounds once strangers (other people's friends) start saving. Real work, real network.

Ready to see your network as a map?

Sign up and add the four people you'd already call back. The graph builds itself from there.

Get started