Solution: A reconciliation engine specialized for messy data: (1) Protocol-aware transaction classification — recognizes LP deposits/withdrawals, perps, restaking, lending, bridging, and wrapping, and applies the correct tax treatment; (2) Defunct-exchange handler — historical price oracles for delisted tokens and dead exchanges so basis can be estimated defensibly; (3) Spam/dust filter — automatically identifies and excludes scammer airdrops that inflate transaction counts; (4) Cross-chain basis tracing — follows assets through bridges and wraps to preserve basis lineage; (5) High-volume optimization — handles 100K+ transactions without manual cleanup; (6) Confidence scoring — flags which transactions have solid basis vs. estimated basis so the user knows their audit exposure. ICP: Active DeFi users and high-volume traders aged 25–45 with thousands to hundreds of thousands of transactions across Ethereum, Solana, L2s, and DeFi protocols. They've bought shitcoins on Chinese exchanges that no longer exist, provided liquidity, traded perps, restaked, and received spam-token airdrops. When they connect their wallets to Koinly or CoinTracker, the tool produces garbage because it can't classify protocol interactions, can't price defunct tokens, and can't trace basis across chains.

Protocol-aware classification and defunct-exchange price reconstruction are the deep technical moats. Mainstream tools treat DeFi as an afterthought; this engine is architected for it from the ground up. The confidence-scoring feature is unique and valuable — it tells users exactly where their basis is solid vs. estimated, which directly informs audit risk and CP2000 exposure.

“Similar to how Plaid built specialized infrastructure for the messy problem of connecting to thousands of banks with inconsistent data formats. This is Plaid-grade reconciliation infrastructure for messy crypto data: the hard technical layer that mainstream tax tools can't match because they optimized for the clean-data majority.”