Settle up, on-chain.
Mend is a non-custodial primitive for shared expenses. Members deploy a group, record expenses against it, and settle the running balance in USDC directly between their wallets — atomically, on-chain, with no third party in the middle.
Status: M1 deployed to Optimism Sepolia. MendFactory at 0x7c6c933b036fce0d6663ab4f3866acdc2a5091da. See docs/specs.md for the full specification and docs/design.md for design rationale.
Splitting expenses with people you trust shouldn't be hard. But it is. Today's tools — Splitwise and its peers — get the tracking part right, but the moment of settlement always falls back to "I sent you the Venmo, trust me." The ledger lives in one app, the money lives in another, and the gap between them is filled with friction: being methodical, depending on others to log their share, chasing settlements that may or may not happen.
Mend closes that gap. Balances and settlement live in the same place, on-chain, and "did you actually pay" stops being a question of trust. The technology is incidental — if a different stack could deliver less friction tomorrow, that would be the right tool. Today, smart contracts on a low-cost L2 are what make this possible: cheap enough to use casually, secure enough to handle real money, and finally mature enough that the ecosystem is no longer hostile to non-technical users.
Mend is built with the conviction that digital products that touch money — and digital products in general — deserve to be made with care. Care in the engineering: formal specs, invariant-tested contracts, a design that earns trust by construction rather than asking for it. Care in the product: language designed for humans rather than the crypto subculture, flows that hide complexity instead of celebrating it, visual quality that signals seriousness. The current state of the on-chain Splitwise space is a handful of bootcamp prototypes and weekend experiments. Mend exists because the problem deserves a serious attempt.
The goal is for Mend to be used. Not as a proof of concept, not as a portfolio piece — as a real tool that real people use to settle expenses with the people in their life. M1 is the foundation: a minimal, audited primitive deployed to Optimism Sepolia, the first step toward a product that hides its plumbing and focuses on the experience of getting things settled. Whether Mend grows beyond that is to be discovered — but the foundation is being built as if it will.
- Two people deploy a Mend group together — one smart contract, one address, just for them.
- Each person does a one-time USDC approval, granting the contract permission to move up to a chosen amount from their wallet (e.g., 1,000 USDC).
- Either person can record a shared expense at any time. The contract updates a single net balance. No money moves yet.
- Expenses can be edited or deleted. The contract recomputes the balance accordingly. A full audit trail is preserved on-chain.
- When the debtor wants to settle up, they call
settle(). The contract pulls the owed amount from the debtor's wallet to the creditor's wallet in a single transaction, and resets the balance to zero.
Directional, not committed. M1 is the current focus; everything beyond it is exploratory and may change, be reordered, or dropped entirely.
| Milestone | Theme | Status |
|---|---|---|
| M1 | Two-party non-custodial IOU contract | Deployed to Optimism Sepolia |
| M2 | Onboarding — embedded smart-account auth, gasless UX | Exploratory |
| M3 | Multi-party groups and debt graph simplification | Exploratory |
| M4 | Off-chain integration — bank-feed ingestion, auto-classification | Speculative |
See docs/design.md for the reasoning behind the milestone ordering.
Built by Ariel Diaz, formerly Smart Contract Engineer at OP Labs (Optimism).
MIT