> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cryptocheckout.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Distribution and claiming
> How the pool splits, why merchants keep 98.505%, and the permissionless property that must not be traded away.
## The split
`distribute()` reads the pool's current token balance and credits each committed recipient to a pull-claim ledger.
```
gross
├── 99.00% merchant
├── 0.75% treasury
└── 0.25% referral partner
```
The percentages are configurable per merchant at pool creation. Only the basis-point **caps** are hardcoded. Production currently runs two configurations: four pools at 100/0 where treasury takes the full 1%, and two at 75/25.
## Merchants keep 98.505%, not 99%
`MerchantPool.distribute()` carves a **50 bps distributor reward from the merchant share** (`MerchantPool.sol:203`), and every live pool carries `distribution_incentive_bps = 50`. The all-in merchant cost is **1.495%**, not 1%.
The site previously advertised "keep 99% / 1% flat, all-in". That was corrected in the 2026-07-28 production-readiness audit, and `src/lib/__tests__/feeDisclosure.test.ts` now re-implements the split and asserts the published copy against it.
The reward exists to incentivise anyone to call `distribute()`. Whether to keep, cut, or re-frame it is an open pricing decision. Note that zeroing it removes the only channel by which the keeper currently recovers deposit-rail gas — see [Gas reimbursement](/economics/gas-reimbursement).
## Permissionless, and why that matters
`distribute()` can be called by anyone. The settlement spec names this **the regulatory linchpin**: because we cannot withhold a merchant's funds, we are not a custodian of them in any meaningful sense.
Any design that gives the platform the ability to block, delay, or condition a merchant's payout destroys this property. It is not tradeable for operational convenience, fee recovery, or debt collection. A claim-gating proposal was evaluated in full and rejected on exactly this ground.
The practical consequence: mechanisms that recover money from a merchant must be **bounded fees inside the split**, never gates on the payout.
## Pull, not push
`distribute()` credits a ledger; `claim()` moves the funds. Separating them means a blacklisted or reverting recipient cannot block the whole distribution for everyone else.
## Who calls it
| Caller | When | Pays gas |
| --------------- | ------------------------------ | ---------- |
| Merchant | Presses Claim in the dashboard | Merchant |
| Keeper backstop | Conditional, monthly | Platform |
| Anyone | Any time, permissionlessly | Themselves |
The merchant path is being consolidated into a single Claim button backed by a `DistributeAndClaim` helper, so one signature does both and the second pool call hits a warm implementation.
## The backstop
`claimable[token][treasury]` is populated **only** by `distribute()`. If a merchant never claims, the platform's 0.75% is never credited and there is nothing to auto-claim — a churned merchant means permanently unrealised revenue and funds sitting in a pool looking like a bug. "Auto-claim without auto-distribute" is not a valid design; the claim would revert.
The eager five-minute cron is being retired in favour of a **monthly conditional** job that distributes a dormant pool and claims only the treasury's share, never the merchant's.
On production today, **11 of 14 real distributions came from the backstop**. This path is load-bearing right now, not theoretical.
Two things change its design once merchants bear deposit-rail gas:
* The backstop becomes the platform's **gas-debt collection path**, not just revenue insurance. Its trigger should be economic — distribute when `treasuryBps × undistributed + gas_owed > distribute_cost` — rather than a pure value threshold.
* The monthly window **is** the gas float. Stretching the cadence from five minutes to 31 days stretches the platform's native-gas exposure by the same factor. The cadence decision and the treasury exposure are one decision, and they were originally written as two independent issues.
## Cost of distributing
Measured 150,516 gas for distribute plus claim on Base. That is roughly \$0.10 per month for 50 merchants — negligible on EVM.
On TRON it is not negligible. A merchant's distribute-plus-claim costs about \*\*$4.59** in TRX, because they sign from their own wallet and therefore burn energy at 100 sun rather than using the platform's rented energy. It amortises over the batch — about $0.23 per order at 20 orders — which makes claim **cadence** a genuine merchant-facing economic decision rather than a preference.