SPV (Simplified Payment Verification)

SPV - Simplified Payment Verification - is the lightweight client model described in section 8 of the Bitcoin whitepaper. It lets a wallet verify its own transactions without storing the full chain.

How it works:

1. The wallet downloads just the 80-byte block headers for every block from genesis to now. That's ~4 MB per year, totaling well under 100 MB after sixteen years.
2. The wallet validates the chain of headers by checking each one's proof-of-work and that each links correctly to its predecessor.
3. For a specific transaction the wallet cares about, it asks a full node for a Merkle proof - a short list of hashes proving the transaction is included in some block's Merkle tree.
4. The wallet recomputes the Merkle root from the transaction and the proof, and checks it matches the root stored in that block's header.

If the proof checks out, the transaction is provably in the chain at the depth specified. The wallet has confirmed inclusion without downloading the rest of the block's data.

What SPV trades off:

• You learn nothing about the rest of the block. You can't verify that other transactions in the block obey consensus rules; you have to trust that the miners and full nodes you're talking to aren't lying about the chain's validity.
• Privacy leakage. Naive SPV asks "what's the proof for transaction X?" - which tells the queried node that you care about X. Better protocols (BIP-157/158 compact block filters, implemented in production as Neutrino) let clients download filters and check for matches locally without revealing which addresses are theirs.
• Eclipse-attack vulnerability. If an attacker isolates your SPV client from honest peers, they can feed you a fork. A full node is more resistant because it independently validates every rule.

Most lightweight mobile wallets use SPV-style logic, often combined with compact block filters for privacy. It's a reasonable tradeoff for everyday use. For high-value or high-stakes use, run a full node.