Rare Pepes don't live on a sidechain. They don't live on Ethereum. They live on Bitcoin, in actual Bitcoin transactions, validated by Bitcoin miners. The way they get there is Counterparty, a metaprotocol that has been running quietly since January 2014.

Most people who talk about Bitcoin NFTs in 2026 mean Ordinals or Stamps. Counterparty predates both by nearly a decade and works on a fundamentally different architecture. Understanding that architecture is how you understand why a token issued in 2016 still exists, still transfers, and still trades on the same chain it was minted on, without any bridge or sidechain standing between it and Bitcoin's proof-of-work.

What a metaprotocol actually is

The Counterparty whitepaper distinguishes between three approaches to extending Bitcoin: overlay networks, colored coins, and metaprotocols.

A colored coin protocol like Ordinals uses Bitcoin's native scripting and UTXO system. It's tied to what Bitcoin Script can do, which is basic token creation and transfer.

A sidechain like Liquid, Rootstock, or Stacks runs an independent blockchain and uses a two-way peg to move Bitcoin in and out. It has its own consensus mechanism and its own validators. Bitcoin's security guarantees apply only insofar as the peg holds and the sidechain validators behave.

A metaprotocol is different. Counterparty doesn't have its own chain. It encodes additional protocol messages inside Bitcoin transactions that regular Bitcoin nodes ignore. Counterparty nodes parse those messages, validate them against Counterparty's rules, and derive a state machine on top of Bitcoin's state machine. Every Counterparty transaction is also a real Bitcoin transaction. Bitcoin miners secure both.

The practical consequence: there is no Counterparty network to attack independently. To attack Counterparty's history you would have to attack Bitcoin's history.

How an asset gets issued

Issuing a Counterparty asset is a Bitcoin transaction. You broadcast a transaction with a dust output back to yourself and an OP_RETURN output containing the Counterparty-encoded issuance message. The message specifies the asset name, supply, divisibility, and a few other parameters. You pay the standard Bitcoin transaction fee, plus a small XCP fee for the asset name if you want a human-readable name rather than a numeric identifier.

Once that transaction is mined into a Bitcoin block, every Counterparty node sees it, parses the OP_RETURN data, and records the new asset in its local state. From that moment, the asset exists. Anyone running Counterparty can verify the issuance by reading the Bitcoin blockchain.

For a Rare Pepe, this is the entire on-chain footprint. The asset exists, it has a name (NAKAMOTOCARD, HOMERPEPE, PEPECASH), it has a supply, and it has an owner. The image itself lives off-chain. The token references it.

This is the pattern every NFT standard later codified: token on-chain, image off-chain, hash for integrity. Counterparty did it first in 2014, using a richer state machine than Bitcoin Script's, and the basic primitive (data inside an OP_RETURN) is the same one Ordinals later used for inscriptions, with different tradeoffs about what to store and where.

XCP and what it does

Counterparty has its own native currency: XCP. It exists because Bitcoin can't directly act as a working currency for Counterparty operations. Bitcoin miners aren't aware of Counterparty, so they can only price transactions by their byte size. XCP gets used wherever Counterparty itself needs to charge for something, most notably named asset issuances and certain on-chain operations.

The initial XCP distribution was unusual for 2014. Instead of an ICO, Counterparty used "proof of burn." For one month in early 2014, users sent Bitcoin to a verifiably unspendable address and received XCP in return. Approximately 2,140 BTC were burned in the process. The XCP supply is fixed at that distribution.

For Rare Pepes specifically, XCP shows up in two places. First, the issuer of any named asset (most Rare Pepes have human-readable names) had to burn a small amount of XCP for the name. Second, until PEPECASH took over as the marketplace currency in September 2016, XCP was a common pairing on the decentralized exchange.

The decentralized exchange

Counterparty has had an on-chain decentralized exchange since 2014. It matches orders entirely through Bitcoin transactions: you broadcast a transaction containing a buy or sell order with terms, and the DEX state machine matches it against existing orders.

This is how the first Rare Pepe traded. Mike's listing for the Satoshi Nakamoto card was a standard Counterparty DEX order pointing at his newly issued asset. Anyone running Counterparty could see the order. Anyone with the matching asset or willingness to pay could fill it.

The DEX has tradeoffs. It's slow by 2026 standards because orders settle in Bitcoin block time. It has limited liquidity compared to centralized order books. But it's permissionless, custody-free, and doesn't require trusting a marketplace operator. Today most Rare Pepe trading happens on higher-level marketplaces that interact with the DEX or use Counterparty's transfer functionality directly, but the underlying capability has been continuously available for over a decade.

Writing in the margins

The Counterparty whitepaper uses a phrase for its design approach: writing in the margins of Bitcoin transactions. The metaphor is precise. Bitcoin has spare bytes in its transaction format (OP_RETURN outputs, witness data) that aren't strictly required for basic transfer semantics. Counterparty uses those bytes to encode protocol messages without modifying anything Bitcoin actually validates.

This is what gives Counterparty its longevity. Every Counterparty operation is a valid Bitcoin transaction. Bitcoin nodes don't need to be aware of Counterparty for Counterparty to work. If every current Counterparty implementation disappeared tomorrow, the data would still be in the Bitcoin blockchain, recoverable by anyone willing to write a parser.

The implication for Rare Pepes is straightforward. The cards minted in 2016 are not held in trust by a foundation, not secured by a third-party indexer, not dependent on a particular hosting service. They are in Bitcoin transactions on the Bitcoin blockchain. As long as Bitcoin works, they exist.

What this means in practice

Three things follow from this architecture for anyone trying to understand Rare Pepes.

First, the cards from 2016 are still on the same chain they were minted on. There is no migration, no version 2, no bridge to maintain. The data is exactly where it was put.

Second, the security model is Bitcoin's security model. The Rare Pepe community didn't have to bootstrap their own validator set. They inherited Bitcoin's.

Third, Counterparty's design from 2014 anticipated the structural problem every Bitcoin NFT protocol later faced: how do you put complex assets on a chain whose scripting language can't natively represent them? Counterparty's answer was the metaprotocol pattern. Ordinals, Runes, BRC-20, SRC-20, and Stamps are all variations on the same idea, arrived at later, with different tradeoffs. The pattern is older than most of the people now debating it.

The technical underpinning is what made Rare Pepes possible in 2016, and it's why they still work today. The same architecture is what lets a modern bitcoin nft marketplace settle trades directly on the Bitcoin chain without bridges or custodial intermediaries.