Ordinal Sats

The sat family: inscribed sats, bitmaps, sat names and CAT-21 sats. How ord binds files to sats, how BTI books them, and the family's live share.

Ordinal Sats 7.22 % of the chain's bytes in Book A, as of block 972,299, 2026-10-20 09:30:48 UTC

Ordinal Sats is the display name of the sat family on block.space. It was renamed from "Individual Sats" on 2026-10-19, and the key stays sat in every URL, API field and database row. BTI's token table calls the sat "the base token": every sat is separately ownable and moves with its output, and its identity is ord's numbering, the ordinal number.

How the data reaches the chain #

Most of the family's bytes are inscriptions. ord binds content to a sat through a witness envelope: an OP_FALSE OP_IF … OP_ENDIF block inside a taproot script-path spend, marked 6f7264 ("ord"). The data-types reference gives the reasons the channel won. Each data push is capped at 520 bytes, so content is chunked, but a tapscript leaf has no script-size limit, so one leaf can hold megabytes. Witness bytes weigh one weight unit each, a quarter of base bytes. An inscription takes two transactions: a commit and a reveal.

CAT-21 uses no envelope. Its mint is any transaction with nLockTime = 21, a field every transaction already has.

What BTI counts as the token #

Under ruleset v1's rule U3, a token is "exactly what can be transferred by itself, proven on chain". An inscription is content, not a token: the inscribed sat is the token that claims the bytes (case law B14). The family has these parts in the token table and on block.space:

sub-item	what narrows the sat	catalogued by
Ordinal Inscriptions	ord bound content to it	the ord adapter (module inscriptions)
Bitmaps	the inscription is a district claim; the district number is a label	the opi-bitmap adapter
Sat names	the inscription names the sat; the name is a label	the opi-sns adapter
CAT-21 sats	an `nLockTime = 21` mint; the cat is a label on the first sat of output 0	the ord adapter (module cat21)

Rarity is not a sub-item. The owner ruled on 2026-08-02 that rarity is a property of a sat, not another token. Uninscribed sats would arrive with the positional appendix, which the token table lists as deferred.

The ord adapter's trust is "definitional" in the trust triad: ord's ruleset is the protocol for which sat an inscription is bound to. A second-source adapter, opi-ord, sights the same sats through OPI. It writes no bytes and cannot create a token. As everywhere in BTI, the bytes come from the raw chain, never from ord.

How the bytes are booked #

An inscription reveal is usually dedicated to its sat, and the sat's claim covers the whole transaction. The commit that enabled it is an enabler: its bytes join the pool and reach Book A only. Case 7 of the BTI case studies is a 3.9 MB jpeg at block 960,962. Its body is 3,888,767 bytes (Book C), the reveal is 3,911,459 bytes (Book B), and the 419-byte commit brings Book A to 3,911,878.

A sat often shares a reveal with other tokens. In case 10, the first BRC-20 mint, the sat and the tick ordi divide the envelope they share. The sat takes 284.5 bytes of Book B and all 71 bytes of Book C, because the mint JSON is the sat's content. Case 3 adds a rune on the same reveal, so three tokens share it.

Sending an inscribed sat is a sighting. ord records the new location, but the byte difference is zero, so the send is judged carriage and adds nothing to the books (U5). A marketplace sale has no record in the protocol, so it stays unattributed, an understatement the ruleset accepts on purpose.

The largest inscriptions ranks inscriptions by Book C, the file's own bytes. A delegate inscription is ranked by the few hundred bytes it bought, not the content it shows.

Live figures #

The family's books (as of block 972,299, 2026-10-20 09:30:48 UTC):

measure	value
Book A	56.13 GB
Book C	33.44 GB
share of the chain's bytes (Book A)	7.22 %
share of all Book A bytes	41.97 %
tokens (sats with claims)	125,519,930
operations	129,015,098
first height	767430

The protocols API splits the family by claim source. Read on 2026-10-20 (books at block 972,199), the inscriptions source covered 125,448,674 sats, sat names 4,956,536, bitmaps 972,116 and CAT-21 63,668. The sub-items overlap, because one sat can be sighted by several sources, so they do not add up to the family total.

ord itself counts 128,169,591 inscriptions, 472,043 of them cursed (as of block 972,308, 2026-10-20 11:17:25 UTC). On 2026-10-19, 582 of the 100,000 largest inscriptions had no sat binding below the export ceiling, which block.space shows as "not yet attributed".

Holders: read on 2026-10-20, the leaderboard's coverage table placed 99.88 % of the family's Book A on 3,034,207 addresses (see holders).

Measured findings #

In the July 2026 measurement run, inscriptions had a bytes:vbytes ratio of 3.374. The measurements page reads this as paying for about 30 % of what they consume, the largest witness discount measured. The same run found that data payload peaked at 45.4 % of block bytes at height 833,000, during the inscription era, and fell to 31.9 % by 850,000 after runes moved to OP_RETURN.

Sources #

Live figures: ord as of block 972,308, 2026-10-20 11:17:25 UTC; protocols as of block 972,299, 2026-10-20 09:30:48 UTC. Refreshed hourly.