A discontinued code prefix keeps its entire back catalogue but stops receiving new serial numbers. That happens for two quite different reasons: the label rotated to a successor prefix and is still active, or the label itself stopped releasing. Telling the two apart is the practical question, because only the first case gives you somewhere else to look.
Nothing about retirement invalidates existing codes. SSNI-500 identifies its
release exactly as it did on the day it shipped. What retirement removes is the
future — a serial above the last one issued under that prefix was never
allocated and never will be.
What is the difference between a retired prefix and a closed studio?
A retired prefix has a successor; a closed studio does not.
| Retired prefix, active label | Label no longer releasing | |
|---|---|---|
| New releases | Yes, under a new prefix | No |
| Back catalogue codes | Valid | Valid |
| Successor prefix exists | Yes | No |
| Catalogue listings | Maintained | Gradually pruned |
| Best search strategy | Search the successor line | Search archives and databases |
The second column degrades over time in a way the first does not. An active label has commercial reasons to keep its back catalogue listed; a dormant one does not, so its titles quietly drop out of storefronts and survive mainly in databases and secondary listings.
Which prefixes were retired but replaced?
The largest labels have all rotated at least once, and in each case the successor is still receiving numbers.
| Retired prefixes | Successor | Label | Successor opened |
|---|---|---|---|
SOE, SNIS, SSNI, SSIS, SONE |
SNOS |
S1 No.1 Style | Nov 2025 |
IPTD, IPZ, IPX |
IPZZ |
IdeaPocket | Mar 2023 |
MIDD, MIDE, MIDV |
MIDA |
MOODYZ | Feb 2025 |
ABP, ABW |
ABF |
PRESTIGE | Jul 2023 |
JUX, JUY, JUL, JUQ |
JUR |
Madonna | Dec 2024 |
JUFD |
JUFE |
Fitch | Jan 2019 |
KAWD, CAWD |
CAWB |
kawaii* | Jul 2026 |
EBOD |
EBWH |
E-BODY | Jul 2023 |
PPPD |
PPPE |
OPPAI | Jan 2022 |
DASD, DASS |
DSOD |
Das! | Aug 2026 |
MEYD |
MFYD |
Tameike Goro | May 2025 |
HND |
HMN |
Honnaka | Jun 2021 |
WANZ |
WAAA |
Wanz Factory | Nov 2020 |
RBD |
RBK |
Attackers | Jan 2021 |
STARS |
START |
SOD Create | Mar 2024 |
NHDTA, NHDTB |
NHDTC |
Natural High | Jan 2025 |
NSPS |
NSFS |
Nagae Style | — |
DVDMS |
DVMM |
Deep's | — |
FSDSS |
FNS |
FALENO | — |
VENU |
VENX |
Venus | — |
RCT |
RCTD |
ROCKET | — |
Dates are the month the successor's first serial shipped, taken from each
label's own catalogue and from FANZA product records, checked 2026-08-03. A dash
means we could not establish the month. HND → HMN and STARS → START are
both genuine successions rather than parallel lines: HND-999 and HMN-001
carry the same release date, and START-001 shipped a month after STARS
closed at 999, under the same label name.
Reading this table backwards is often the fastest way to recover from a failed lookup. If a serial under a retired prefix is out of range, the same performer, series or concept usually continues under the successor, and the successor's serial numbering restarted near 001.
Which prefixes have no successor?
Fewer than published lists claim. Every prefix in this group that we checked turned out to have a live continuation somewhere, which is worth knowing because these are exactly the ones commonly listed as dead labels.
| Prefix | Label | What we found |
|---|---|---|
DV |
Alice Japan | The label's DVAJ line was still receiving numbers in mid-2026 |
WANZ |
Wanz Factory | Reopened as WAAA in November 2020 and still active |
DVDES |
Deep's | Ran alongside DVDMS, which closed in 2022; Deep's now numbers DVMM |
RBD |
Attackers | Succeeded by RBK, which opened a month after RBD closed |
We are deliberately cautious in this section. "The prefix stopped receiving numbers" is something you can observe from release listings, and it is all we claim above. "The company closed" is a corporate fact that requires a source, and English-language coverage of Japanese adult industry corporate changes is thin and frequently wrong. Where we could not confirm the second, we state only the first.
Why do older codes get harder to find over time?
Because catalogues prune, licences lapse, and storefronts close — while the code itself stays perfectly valid.
Four effects compound:
- Storefront pruning. Inactive listings are removed to keep catalogues navigable, so the authoritative record disappears first.
- Licence expiry. Distribution rights for older material lapse and are not always renewed, particularly for labels that no longer negotiate.
- Format obsolescence. Titles released only on physical media, or in resolutions no longer offered, are the first to be dropped from digital catalogues.
- Database drift. Secondary databases inherit gaps from the storefronts they scraped, so a gap in one propagates.
None of these make the code wrong. They make the index incomplete, which is a different failure and calls for a different response — see JAV code not found.
How do you search a code from a discontinued prefix?
Work outward from the code rather than reformulating it.
- Confirm the serial is in range. A serial above the prefix's last issued number cannot exist. If you are guessing at a number, this is usually why.
- Check the successor prefix using the table above and the studio code reference.
- Search databases rather than storefronts. For dormant labels the catalogue record often outlives the sales listing.
- Identify the performer instead of the title. A named performer's filmography is a more durable index than any single storefront.
- Do not add suffixes. Old titles attract
-C,-Uand-4Kvariants in circulation, and none of them are indexed anywhere.
Step 4 is the one that reliably works when the others do not, because a performer's identity does not get pruned from a catalogue the way a listing does.
What do you do when the catalogue trail has gone cold entirely?
Match the face, since it is the identifier that survives delisting.
Discontinued material is precisely the case where text search runs out: the storefront is gone, the database entry is thin, and the file name has been rewritten several times. A frame still contains the one thing that has not changed. Our index encodes faces as 512-dimension ArcFace vectors and compares them across 241,792 indexed faces covering 2,333 named performers (our index, 2026-08 snapshot), at a 0.40 similarity threshold calibrated for video stills.
Boundary worth stating: older material is where our coverage is weakest. The index is built from what currently circulates on 106 sites, 2,246 of 2,333 representative images come from a single source, and a large portion of the faces we hold have no name attached to them at all. For a title from a long-dormant label, a confident match is the exception rather than the rule, and we would rather say that plainly than imply otherwise.