Skip to content
LearnBitcoin

LearnGlossary

Inscriptions

A technique for embedding arbitrary data - text, images, code - inside Bitcoin transactions by writing into Taproot witness data.

An inscription is a chunk of arbitrary data - a string of text, a PNG, a piece of code, anything - written into a Bitcoin transaction's witness data using a specific Tapscript pattern. Once inscribed, the data lives in the blockchain permanently, just like any other transaction history.

Casey Rodarmor introduced the technique as part of the Ordinals protocol. The first inscription was confirmed in block 767,430 on December 14, 2022, and inscribing caught on in the first months of 2023. The technique does not require any change to Bitcoin's consensus rules - it works within existing Taproot functionality activated in 2021.

How it works

The inscription is encoded inside a Taproot script using a pattern that looks roughly like this (the notation follows the Ordinals documentation):

OP_FALSE
OP_IF
  OP_PUSH "ord"         // protocol identifier
  OP_PUSH 1             // the next push is the content type
  OP_PUSH "image/png"   // MIME type
  OP_PUSH 0             // everything after this is the content
  OP_PUSH <bytes>       // the data, split into pushes of at most 520 bytes
OP_ENDIF

Because OP_FALSE leaves a false value for OP_IF to test, the script always skips ahead to OP_ENDIF, so the data in between is never evaluated as script - it just sits in the witness. This pattern is sometimes called the envelope. A Taproot script only appears on-chain when its output is spent, so inscribing takes two transactions: a commit that creates an output committing to the script, and a reveal that spends it and exposes the data.

The crucial property is where this data lives: in the witness, not in the transaction's main body. Under SegWit weight rules, each witness byte counts as one weight unit and each non-witness byte counts as four, so data in the witness costs a quarter as much block space. The first inscription's reveal transaction is 1,040 bytes. It weighs 1,286 weight units (about 322 virtual bytes), where it would weigh 4,160 if every byte counted at the full rate. Without that discount, inscriptions would be too expensive to be common.

For the Taproot mechanics underneath, see How Taproot Actually Works.

The community debate

Inscriptions are one of the most contested phenomena in modern Bitcoin culture. The fault line is roughly:

  • Critics argue inscriptions are blockchain spam - they fill blocks with non-monetary data, bloat the UTXO set with small outputs, increase storage and bandwidth requirements for node operators, and push transaction fees up for people trying to use Bitcoin as money.
  • Proponents argue that inscriptions are paying-customer use of block space, that miners are free to mine what users pay for, and that any attempt to restrict valid transactions amounts to censorship of the protocol's neutrality.

Both positions are internally consistent. The disagreement is over what Bitcoin is for.

The debate produced BIP-110, the Reduced Data Temporary Softfork (RDTS). It proposed a one-year soft fork that would cap most data pushes at 256 bytes and make invalid any Tapscript that executes OP_IF, which the inscription envelope relies on. Coins created before activation would have been exempt.

Miner signaling for BIP-110 never reached its 55% threshold. Its mandatory signaling period began at block 961,632 in August 2026. From then on, nodes enforcing BIP-110 rejected every block that did not signal support, which meant rejecting the main chain, and they split off onto a chain that stalled. The BIP was marked Closed in August 2026.

In September 2026, Bitcoin Knots, the node software that had shipped BIP-110, released a hard fork for that minority chain. It swapped Bitcoin's proof of work for a different algorithm, BLAKE2b, and turned on BIP-110-style data limits there. As of October 2026, the main chain had not adopted those rules.

Why it matters

For a Bitcoin learner, inscriptions matter for three reasons:

  1. They're a real change in how the blockchain is used. Whether you welcome the change or oppose it, ignoring it leaves you unable to read fee charts or block contents from 2023 on.
  2. They surface a deep question about Bitcoin's purpose. "Monetary network only" versus "credibly neutral platform" is a live disagreement among serious Bitcoiners.
  3. They reshaped the fee market in 2023. During a wave of text inscriptions in May 2023, Glassnode estimated that those inscriptions paid between 30% and 60% of all transaction fees.

The technique itself is permanent - Taproot is unlikely to be rolled back, and as of October 2026 the envelope pattern is valid under Bitcoin's consensus rules. What can change is the limits placed on what inscriptions are allowed, and that's the live debate.

Key takeaways

  • Data is wrapped in an OP_FALSE OP_IF tapscript envelope and stored as witness data
  • Witness data is fee-discounted, which is what makes large inscriptions economically practical
  • The technique was popularized by the Ordinals protocol in early 2023 and drove major fee-market activity that year

On the mailing list

The topic timeline: Data on the chain: OP_RETURN, inscriptions and spam (107 threads).

Corrected

  • Was: The entry said Casey Rodarmor introduced inscriptions in January 2023, and that inscription demand has been a major driver of fee levels since 2023.

    Now: The first inscription was confirmed in block 767,430 on December 14, 2022, and inscribing caught on in the first months of 2023. The fee effect is dated to 2023, when Glassnode estimated that text inscriptions paid between 30% and 60% of all transaction fees during the May wave. Details

  • Was: The entry listed a Wikipedia article on the Ordinals protocol as an external reference. That article does not exist, and we found no record that it ever did.

    Now: The reference is the inscriptions chapter of the Ordinal Theory Handbook. Details

Sources (8)

External references (1)

Related terms (6)

Published . Revised . Two corrections logged for this page.

Edit this page on GitHub