SDK Adapters

The Yield Engine discovers positions across 15 chains. SDK adapters are the modules that do the actual contract reads — each one knows its protocol's quirks, correct APY fields, and data traps.

How adapters work

Each SDK adapter is a standalone Python module that queries protocol smart contracts directly via JSON-RPC. No pip dependencies -- stdlib only. No cached APIs, no third-party aggregator data. The number you see is computed from the contract state that exists right now.

Protocol coverage

APY field mapping

Every protocol reports APY differently. Adapters normalize everything into a single effective_apy column:

Adapter anatomy

Every adapter follows the same structure:

class AaveAdapter:
    """Direct contract reads for Aave V3 markets."""

    POOL_ADDRESS = "0x87870Bca3F3fD6335C3F4ce8392D69350B4fA4E2"
    CHAINS = ["ethereum", "arbitrum", "base", "optimism", "polygon"]

    def fetch_pools(self, chain: str) -> list[dict]:
        # 1. Multicall getReservesList()
        # 2. For each reserve: getReserveData()
        # 3. Compute supply_apy from liquidity rate
        # 4. Return normalized pool objects
        ...

    def compute_apy(self, liquidity_rate: int) -> float:
        # RAY = 10**27, convert on-chain rate to annualized %
        return ((1 + liquidity_rate / 10**27) ** 31536000 - 1) * 100

Adapters produce raw position data. Before it reaches your board, every position goes through the verification pipeline — exploit cross-reference, TVL checks, APY consistency, and contract verification.