An honest take on why Splitwise, Tricount, and Kittysplit each fall short - and what EvenRound does differently.
There are already three good expense splitters. We built a fourth anyway. Here's why - and what we did differently.
For the last decade, splitting a dinner with friends has meant one of three apps. Splitwise is the household name - it worked, it scaled, and at some point it slowly turned into a subscription product, capping the free tier at four expenses per day in late 2024. Tricount was the European alternative - cleaner, no signup, beloved across DE/FR - until N26 acquired it and development effectively stopped. Kittysplit, the beautiful minimal one, never expanded past its original scope.
None of them are bad. Each one made deliberate decisions, lived with them, and ran into the limits of those decisions. The question we kept asking ourselves was: what would a 2026-native expense splitter look like, given everything that's changed about how groups actually coordinate money?
Three things, specifically:
Most splitting happens in groups where one or two people are enthusiastic and the rest just want to know what they owe. Asking everyone to make an account - verify their email, pick a password, download an app - is a way to ensure half the group never logs in. The person who's tracking ends up doing data entry for everyone else.
Kittysplit got this right ten years ago. We adopted the same model: a group is a shareable URL. Each member gets a magic link they click once. No signup. No password reset emails. Nothing to install.
Restaurant bills are where every splitter gets awkward. Equal split feels unfair when one person had a salad and another had the wine pairing. Itemising by hand is tedious. Splitwise paywalled this. We decided it should be free and instant.
Modern multimodal models can read a receipt and produce structured line items in two seconds. We snap, we extract, you tap names against items. Tax and tip get split proportionally. The same flow that took five awkward minutes in Splitwise takes about thirty seconds in EvenRound.
If five people in a group all owe each other small amounts, the naive result is a tangled web of transfers. Splitwise's "Simplify debts" feature solved this with greedy minimization, and it's the single feature most people stay for. We rebuilt it from scratch - same result, same algorithm, but exposed clearly with a one-tap "show me the plan" rather than buried as a setting.
We didn't build a banking layer. There's no EvenRound Visa card. There's no escrow. We're not trying to hold your money - we're trying to figure out who pays whom, and then get out of the way. Pay through whatever rail your group already uses: Stripe, PayPal, Wise, Revolut, cash.
We didn't build social features. There's no feed, no friend graph, no notifications about other people's expenses. Splitting bills is a chore, not a hobby. The job is to be done with it.
EvenRound is free forever for the core flow. A premium tier will eventually exist for power users (long history retention, custom branding, unlimited AI receipt scans), but it'll never paywall basic splitting. We also plan to open-source the settlement engine, the FX-snapshot logic, and the schemas under MIT - the parts of the product that are math, not service.
This week we're launching the marketing site, the comparison pages, and the use-case guides. The core app is already live. Over the next month we'll ship CSV import from Splitwise and Tricount, native iOS and Android wrappers around the PWA, and the first open-source repo.
If you've been frustrated by the existing splitters - particularly the signup walls and paywalls - give EvenRound a try. We'd love to hear what works and what doesn't.
Free forever. No signup. Works in your browser in 30 seconds.