Skip to content

Fees and burns

Every trade pays the 1% Uniswap V3 fee. That fee is the only value extractable from a launch: the WETH buyers pay for tokens stays locked in the position forever. This page covers how the fee splits, and how the protocol’s share funds an automated buyback-and-burn.

Collected fees split first between the creator and the protocol, then the protocol’s share splits again on-chain:

  • Creator keeps 1 − protocolFeeBps of collected fees.
  • Protocol takes protocolFeeBps. Of that, BUYBACK_SHARE_BPS = 3000 (30%) goes to the buyback treasury and the remaining 70% funds operations.

At the current mainnet configuration protocolFeeBps = 5000 (50% creator / 50% protocol), the 1% trade fee breaks down as:

Destination Share of trade volume
Creator 0.50%
Buyback treasury 0.15%
Operations (infra, dev, audit, legal, marketing) 0.35%

Fees from all launched tokens flow into a single pooled treasury: the pool is a platform-wide flywheel, not escrowed per token.

The locker holds the position NFT and exposes claimFees, which collects the V3 fees accrued by a launch’s position and splits them per the rule above. The collect is capped to owed fees: position principal is never moved (invariant L2). Claiming fees can never touch locked liquidity.

The protocol has no native token, takes no team or insider allocation in any launched token, and holds no bought tokens. Everything the treasury buys is burned. There is no treasury bag anyone can sell.

Selection: merit-only, top 20 by fees contributed

Section titled “Selection: merit-only, top 20 by fees contributed”

Each round, the treasury buys and burns the top 20 launched tokens ranked by fees they contributed to the pool, among tokens that pass a qualification gate.

Fees-contributed is the ranking metric because it is the hardest to fake cheaply and the most defensible: a token’s activity funded the pool, so its activity earns its slot. (Ranking by market cap would bias toward already-large tokens; ranking by raw volume invites wash trading.)

A token must pass all of the following to be eligible:

  • Locked liquidity: true by construction for every Merry Men launch (single- sided, permanently-locked V3 liquidity from block one). Proves the token can actually be traded and is not a honeypot.
  • Minimum age: the token must have existed for a set duration, filtering out instant rugs. (Configurable immutable.)
  • Fee-contribution floor: a sustained activity floor over a rolling window, proving real trading rather than a one-time spike. (Configurable immutable.)

Every token bought is burned on execution. Nothing enters a wallet or treasury the team controls. Each buyback is pure, permanent supply reduction: the honest answer to “can the team profit from the tokens it buys?” is no.

Public, predictable buyback transactions are a front-running target. Two on-chain defences are in place:

  1. Per-buy slippage bound. executeRound(uint256[] minOut) takes a minimum-out per buy; the keeper simulates the round and submits tight bounds, so a sandwiched or bad fill reverts rather than executing.
  2. Keeper-gated execution. When a keeper address is configured, only it may call executeRound, so the round is submitted through a private/relayed mempool and never sits in the public pool. The keeper controls timing only: selection is deterministic, so it can never choose which tokens are bought. Setting keeper = 0 leaves execution permissionless and relies on the slippage bound alone.

The core mechanic is locked; a few operational parameters remain configurable:

  • Round cadence: how often a round fires (time-based, or triggered at a treasury threshold). Batched rounds are preferred: cheaper on gas, and each round becomes a visible on-chain burn event.
  • Budget split across the 20: even split (simplest to explain) versus weighting by fee contribution.
  • Threshold values: the concrete numbers for minimum age and the fee floor.

The buyback treasury (BuybackTreasury.sol) is deployed and unaudited. It is immutable and holds protocol revenue, so a third-party audit and bug bounty are strongly recommended before it holds significant value: this is a recommendation, not a hard launch gate.