Ordinals is a Bitcoin protocol that does two related things. First, it assigns a serial number to every individual satoshi in Bitcoin's emission, in chronological order. Second, it uses Bitcoin's witness data to attach arbitrary content (images, text, files) to specific satoshis, calling the result an "inscription." Combined, you get a non-fungible token on Bitcoin where the content is stored directly on the chain.
Casey Rodarmor built the protocol and launched it publicly on January 21, 2023, though the genesis inscription (a black-and-white pixel-art skull, designated inscription #0) had been created on December 14, 2022. Rodarmor's blog post "How Ordinals Came to Be" gives the long account of the development; the short version is that he had been thinking about how to put NFTs on Bitcoin since 2017 and arrived at ordinal theory as the cleanest answer.
Ordinal theory: numbering the satoshis
Bitcoin's smallest unit is the satoshi: one hundred-millionth of a bitcoin. There are approximately 2.1 quadrillion satoshis in Bitcoin's eventual total supply. Each one is, in a sense, distinguishable from the others by where it sits in Bitcoin's emission history.
Ordinal theory formalizes that distinction. It assigns each satoshi a sequence number based on the order in which it was mined, starting from satoshi #0 in the genesis block. The numbering scheme also tracks each satoshi as it moves through Bitcoin transactions, applying a deterministic rule about how multi-input or multi-output transactions split satoshi balances. The result is that any satoshi anywhere in the Bitcoin UTXO set can be identified by its ordinal number.
This is purely a numbering convention on top of Bitcoin. Bitcoin itself doesn't care about ordinal numbers. The protocol doesn't require any consensus changes. Software running ordinal theory just looks at the existing chain and applies the numbering rules to derive the satoshi identities.
From this base, Rodarmor built classifications. Some satoshis are "rare" by virtue of their position: the first satoshi of each block (uncommon), the first of each difficulty adjustment (rare), the first of each halving epoch (epic), the first of each cycle of four halvings (legendary), and the original satoshi #0 (mythic). These rarity tiers are conventions, not protocol facts, but they've shaped how the ordinals community thinks about value.
Inscriptions: content on a satoshi
An inscription is data attached to a specific satoshi. The data lives in Bitcoin's witness segment, which was introduced by SegWit in 2017 and which allows arbitrary bytes in transaction inputs at a discounted weight relative to non-witness data.
Mechanically: someone creates a Bitcoin transaction whose witness data contains content (an image, text, video, code, anything up to roughly 4 MB) plus a content-type marker following the inscription envelope format. The transaction is broadcast and mined. Ordinals indexers (running the ord software or compatible) read that transaction, extract the inscribed content, and associate it with whichever satoshi the transaction's first input was pointing at.
From that point, the inscription "lives on" that satoshi. As long as the satoshi stays in the UTXO set (whoever holds the corresponding private key controls it), the inscription is associated with them. Sending the satoshi sends the inscription.
The content is stored entirely on the Bitcoin blockchain. Unlike Counterparty NFTs, which store metadata on-chain and host the image externally, Ordinals inscriptions have the image, video, or text bytes themselves in Bitcoin's transaction history. As long as Bitcoin exists, the inscription is recoverable.
The technical caveat: witness data is theoretically prunable by Bitcoin Core. A future Core release could decide to discard old witness data after some confirmation depth. In practice, most full nodes retain witness data indefinitely, and the consensus protocol doesn't require pruning. But the architectural difference between "stored where Bitcoin must keep it" (UTXO set, used by Stamps) and "stored where Bitcoin keeps it by convention" (witness data, used by Ordinals) is real.
BRC-20: fungible tokens on top of Ordinals
BRC-20 launched on March 8, 2023, less than two months after the Ordinals protocol itself. The creator, working under the handle "domo," realized that the inscription mechanism could carry not just images and text but structured JSON, and that you could use that JSON to encode a fungible token standard.
The first BRC-20 token was ORDI. The standard works through three operation types encoded as JSON inscriptions: deploy (define a new token), mint (claim a portion of supply), and transfer (send between holders). All three are just inscriptions; the indexer software interprets them to derive token balances.
BRC-20 was both successful and controversial. Successful because it drove inscription volume to historic highs in 2023, with millions of inscriptions in some weeks and a brief period where Bitcoin block space was almost entirely BRC-20 minting activity. Controversial because Rodarmor publicly distanced himself from BRC-20, calling fungible tokens on Bitcoin "99.9% scams and memes" while acknowledging that they were unstoppable.
Runes: the second-generation fungible standard
Runes is Rodarmor's own answer to BRC-20. It was first proposed in a Rodarmor blog post on September 25, 2023, and launched at the Bitcoin halving on April 20, 2024. The design uses Bitcoin's UTXO model directly rather than the inscription mechanism, which means smaller on-chain footprints, fewer junk UTXOs, and what Rodarmor describes as "responsible UTXO management."
Runes is technically distinct from Ordinals; the two protocols share an author and an ecosystem but are separate systems. Most Ordinals marketplaces support Runes trading as a separate category. The market response to Runes has been more measured than the initial BRC-20 boom, partly because the cycle had cooled by mid-2024 and partly because Runes was designed to be more efficient rather than more provocative.
The chain-bloat debate
Ordinals' arrival in January 2023 immediately triggered debate within the Bitcoin developer community. Block sizes spiked. Transaction fees climbed dramatically. Bitcoin Core developer Luke Dashjr publicly argued that inscriptions were "exploiting a vulnerability" in Bitcoin Core rather than using a sanctioned feature, and threatened to fix the loophole in a future release.
The technical specifics of that debate: Bitcoin Core has long had a -datacarriersize setting that limits how much data a transaction can carry in an OP_RETURN. Inscriptions sidestep that limit by encoding data in witness data using a Bitcoin Script pattern, which the Core software treats as program code rather than data. Dashjr's view is that this is data masquerading as code and should be limited.
Rodarmor's response has been mixed. He has acknowledged that inscriptions impose costs on the network and that the concerns about chain growth are legitimate. He has also defended the protocol as making productive use of Bitcoin's existing capabilities. As of 2026, the practical compromise is that inscriptions continue to be allowed by default in Bitcoin Core, the chain-size concerns have not produced a hard fork or filtering change, and the Ordinals ecosystem operates with the implicit understanding that its capabilities depend on Core remaining accommodating.
Ordinals in 2026
After the Magic Eden exit in February 2026, the Ordinals marketplace landscape consolidated around specialist platforms: Satflow, Gamma, UniSat, Ordinals Wallet, and a handful of others. Trading volume is reduced from the 2024 peak but the protocol remains the dominant Bitcoin NFT category by transaction volume, asset count, and cultural attention.
The historical claim Ordinals can defensibly make: it was the first protocol to put substantial digital content directly on the Bitcoin chain in a structured, indexable way. It was not the first bitcoin nft protocol. That's Counterparty, by nine years. But Ordinals brought Bitcoin NFTs into mainstream awareness, drove the development of the modern marketplace ecosystem, and produced the first BTC-native projects to cross meaningful market valuations. The cultural impact has been disproportionate to the technical novelty, which is usually how these things work.