The problem
I cannot hold my own attention, because breaking it costs me nothing in the moment. Tabbing out to check a feed feels free while I am doing it, and the bill only arrives later as work I did not do, so I do it constantly. The tools meant to fix this never changed that maths for me. Timers, site blockers and streak apps have nothing at stake, so ignoring them costs nothing either, and I ignore them.
The only thing that has ever worked on me is knowing someone else can see whether I showed up. Online, that runs into a trust problem the moment real money is involved to make it bite: somebody has to hold the pot and decide who kept their word. If that somebody can keep the money, move it, or pay the wrong people, the incentive that made it work in the first place is gone. What I wanted and could not find was a way to put a real cost on breaking my focus without having to trust whoever holds the money.
My approach
Monk puts the money on Monad and makes the operator a referee that cannot cheat, only judge. Everyone in a room stakes the same amount of MON into a room on the contract. The contract holds every stake, not me. When the session ends, the referee reports who stayed present, and the contract pays the whole pot to those survivors in equal shares and burns the rest.
The design choice that matters is what the referee is allowed to do. The contract enforces three things on chain. The referee can only pay out a room's own players and can pay no one else. It can only settle after the session's clock has actually run out. And if the referee never settles, anyone at all can trigger a full refund to every player after a one hour grace window. So the worst the operator can do is fail, and failing just returns everyone's money. If nobody survives a session, the pot is burned to the dead address rather than kept, so the house takes nothing.
What the chain proves, and what does not
The chain proves custody and the payout rules: that the stakes are escrowed by the contract, that a settlement can only pay the room's own players or burn, and that a stuck room always refunds.
The chain does not watch your screen. Whether a given player stayed present is judged off chain by the referee from presence signals, and that judgement is trusted. What Monk removes is not trust in the judgement of presence, it is trust in who holds the money, which is the part that could take your stake.
What I built
A Solidity contract, Monkmode.sol, deployed and Sourcify verified on Monad testnet at 0xBbE7f526cb9Bf646CcEd56784D817822Cee45634. It escrows each room's stakes, holds the start time and duration, pays survivors in equal shares, and burns the pot when no one survives. Settlement is referee only and refuses to run before the session clock is up, and a permissionless full refund opens after the grace window. It is covered by a Hardhat test suite. A rehearsal script that drives a real room on the deployed contract with a real survivor and a real breaker is in the repo, and it has been run against the deployed contract on testnet, not only against a local fork.
A web app with wallet connection for staking, a presence engine that reads leaving the room tab as breaking focus, daily public pools where the first entrant opens the room, and a Teams mode where the winning side's survivors are passed as the contract's survivor set with no contract change.
A public record derived from the chain. A wallet's history of sessions and its streaks are read from on chain events rather than from my database, so anyone can check a player's record.
What you cannot use yet, since this section is only worth reading if it is exhaustive: presence is read from the room tab and nothing else. Keeping the tab focused suits reading or thinking, but it punishes anyone whose work lives in another application. A browser extension that widens the read to the whole browser is written and is not published, so it is not part of what is in front of you, and it is not counted as shipped until you can install it.
Why it is important
It shows a narrow, honest use of a chain: hold the money and bound the operator, and trust the operator only for the judgement a chain genuinely cannot make. A lot of on chain products claim to remove trust and quietly keep it where it counts. Monk states plainly which part is trustless, the custody and the payout rules, and which part is trusted, the read of who stayed present, and it makes the trusted part safe by ensuring the operator can never do worse than hand everyone their money back.
It also makes group focus work with strangers. Accountability groups usually need people who already know each other, because the pot depends on nobody running off with it. When the contract holds the stakes and the operator is boxed in, a room full of people who have never met can still put real money on holding their attention together, and the worst case is a refund.