100 articles on finding titles, reading codes, platforms and how this index works.
Most failed searches come from a handful of repeatable errors: using near-duplicate image search on a self-captured frame, reading rank order instead of similarity scores, retrying the same bad frame, and treating a performer identification as a title identification.
5 min readFace search converts a face into a list of numbers and compares those numbers instead of comparing pictures. Detection finds the face and five landmarks, alignment straightens it, an encoder turns it into a 512-number vector, and cosine similarity ranks it against every indexed face.
5 min readSearching by name fails more often than searching by face, because one performer can carry a Japanese name, several romanised spellings, and one or more stage aliases. Establish the canonical name form first, then search each variant, then verify identity visually before trusting a filmography.
5 min readA JAV code is a studio label prefix plus a serial number, and it is the most reliable identifier a release has. Normalise the code to LETTERS-DIGITS, search it in that form, then check the alternate forms — zero-padded, unspaced, and suffixed — before concluding it does not exist.
5 min readFinding a video from one screenshot works by identifying the performer's face rather than matching the picture. Detect the face in the frame, convert it to a numerical vector, compare it against an index of faces from adult covers, then confirm the title from the performer's filmography.
5 min readA usable screenshot needs one clearly visible, roughly front-facing face captured at the player's native resolution. Background, lighting and scene content are irrelevant because alignment discards them. Occlusion, extreme profile angles and motion blur are what actually break detection.
5 min readOfficial sources identify their operator, sell through recognised payment processors, carry complete and consistent catalogue metadata, and honour removal requests. Unauthorised sites hide ownership, monetise through intrusive advertising, host inconsistent re-encodes, and offer current releases free in full.
5 min readSynthetic video betrays itself at boundaries and across time. Check hands and finger counts, earrings and jewellery, hair strands against the background, teeth and tongue during speech, and whether details stay consistent from second to second. Single frames deceive; sequences rarely do.
7 min readSafety depends on what actually leaves your device. Most image search tools upload the whole file to a server. A browser-side pipeline sends only a cropped face region and five landmark coordinates, so the original photo, its metadata and everything outside the face never leave your machine.
5 min readImage search tools split into two families that solve different problems. Near-duplicate matchers look for copies of a whole picture in a web crawl. Face-vector engines encode a face and compare identities. Only the second family can work on a frame you captured yourself.
7 min readA removed page usually means one of four things: a copyright takedown, a licensing window ending, a consent-based removal request, or the host itself going offline. Identifying the reason tells you whether the title is legitimately available elsewhere or should not be sought at all.
5 min readA title goes missing for four distinct reasons: the search input is too weak to identify anyone, the performer is in the index but unnamed, the release was never indexed by any source being searched, or the title exists under a different code or translated name than the one being searched.
5 min readGoogle reverse image search fails on adult video screenshots because Google does not crawl most adult sites, and because it matches whole-image composition rather than faces. Finding a performer from a frame requires face-vector matching against an index built specifically from adult covers.
4 min readA discontinued code prefix keeps its entire back catalogue but receives no new serial numbers, which happens either because the label rotated to a successor prefix or because the company stopped releasing. Distinguishing the two cases determines whether a successor prefix exists to search instead.
7 min readFC2 codes identify uploads on a self-publishing marketplace and take the form FC2-PPV-1234567, a single global serial with no label prefix. Studio codes identify catalogue releases from a production company and encode which label issued them, which makes the two families behave differently in every search.
5 min readAmateur and planning-series releases are numbered by series rather than by studio, producing codes such as 259LUXU-1234 where a numeric platform identifier precedes the series tag and a serial follows it. That serial counts entries published in the series, not products issued by a production label, which is why the numbers climb far higher.
6 min readUncensored subscription sites number releases by publication date rather than by label, producing codes such as 010124-001 where the digits are a date and a daily sequence. A few sites use their own sequential serials instead, and neither convention carries a studio prefix.
6 min readA JAV code returns no results for six recurring reasons: a mistyped or mis-transcribed code, leftover suffixes and site tags, an internal storefront ID rather than the printed code, a non-studio numbering family, a retired prefix, or a release the catalogue you searched never carried.
5 min readPrefixes from A to M cover PRESTIGE (ABF, ABW, ABP), Attackers (ADN, ATID), kawaii* (CAWD, KAWD), Das! (DASS, DASD), E-BODY (EBWH, EBOD), FALENO (FSDSS), IdeaPocket (IPZZ, IPX), Madonna (JUQ, JUL), Fitch (JUFE), MOODYZ (MIDV, MIAA) and Tameike Goro (MEYD), each listed below with whether the prefix still receives new numbers.
7 min readPrefixes from N to Z cover Natural High (NHDTC, NHDTB), Nagae Style (NSFS), OPPAI (PPPE, PPPD), Premium (PRED), ROCKET (RCTD), SOD Create (START, SDAB, SDNM), Attackers (RBK, SHKD), S1 No.1 Style (SNOS, SONE, SSIS, SSNI), Venus (VENX), Wanz Factory (WAAA, WANZ) and Max-A (XVSR).
7 min readStudio code prefixes map to the label that issued a release: SNOS and SONE to S1, IPZZ to IdeaPocket, MIDA to MOODYZ, ABF to PRESTIGE, START to SOD Create, MFYD to Tameike Goro. Labels rotate prefixes over time, so most studios own several — one current, the rest retired.
8 min readA JAV code splits into four layers — a label prefix, a serial number, an optional internal padding used by retailers, and optional suffixes added downstream. Reading a code means knowing which layer each character belongs to, because only the first two come from the studio.
7 min readSuffixes attached to a JAV code describe a file rather than a release: -C marks a third-party subtitled copy, -U marks a copy claimed to be uncensored, and -4K claims a resolution. None are issued by the studio, and none are verifiable from the file name alone.
6 min readA JAV code is the catalogue number a Japanese adult video label prints on a release — a letter prefix, a hyphen and a serial number, as in SSIS-001. The prefix identifies the label or series, and the number identifies the individual title inside that series.
6 min readOne release carries several codes because each distribution channel numbers it independently — the retail disc, the digital storefront, a re-release, a VR edition and a compilation disc all get their own catalogue number, even when the footage is identical.
6 min readCensored and uncensored Japanese adult platforms are two separate markets rather than the same films with and without a filter. Japanese law requires genitalia to be obscured in material distributed domestically, so uncensored titles come from producers operating outside Japan, with their own studios, performers, catalogues and subscription sites.
9 min readNeither model is cheaper in general. A FANZA monthly subscription beats single purchases above a break-even volume — the monthly fee divided by the typical per-title price — and loses below it. Volume decides cost, but permanence, catalogue rotation and rewatching decide whether cost is the right question.
8 min readFANZA and MGS are both official Japanese distribution platforms, but they differ in catalogue scope and buying model. FANZA operates as a broad storefront carrying many studios; MGS is a Prestige-group storefront that licenses in many other makers but reserves exclusive and early releases for its own group. The right choice depends on what you are looking for, not which is larger.
12 min readFive checks prevent almost every bad adult-platform subscription: whether your payment instrument will be accepted, whether the service plays in your country, what the plan actually includes, where cancellation happens, and how renewal and refunds work. Each takes about two minutes and is answerable before you pay.
9 min readFree adult sites are rarely free in the sense people assume — the cost is paid in attack surface, through aggressive ad networks, fake player prompts, notification hijacking and account data exposure. Paid platforms shift that risk towards billing instead. Knowing which risk each side carries is what makes the choice a real one.
8 min readCancelling an adult subscription happens wherever the billing relationship lives, not on the site you watched on. Card payments cancel on the platform, app-store purchases cancel in the store, and wallet or carrier billing cancels in that account. Capture a confirmation, then verify against the next statement.
9 min readJAV database sites fall into four structural categories — storefront catalogues, community metadata databases, performer-centric indexes and image or face indexes. Each is authoritative for a different field, and none is complete. Knowing which category you are reading determines how much you should trust a given entry.
8 min readJapanese adult platforms typically accept domestic credit cards, prepaid codes, carrier billing, digital wallets and bank or convenience-store transfers, with international cards the least reliable of the group. Each method differs in overseas availability, in how it fails, and in where you must later go to cancel.
10 min readRegion locks are not one mechanism but six: edge geoblocking by IP, per-title licensing restrictions, payment-instrument country rules, age-verification law in the viewer's own country, network-level filtering by a local ISP, and Japanese-only interfaces. Each produces a different symptom and only some have a legitimate route around them.
9 min readStreaming plays a title from the platform's servers on demand, while downloading stores a local copy that is usually DRM-wrapped and bound to your account rather than owned outright. Streaming suits browsing on a stable connection; downloading suits weak bandwidth, offline viewing and the highest quality tier.
9 min readUncensored Japanese adult platforms differ on ten comparable dimensions — pricing model, archive access, release cadence, exclusivity, downloads, resolution tier, VR, interface language, accepted payment instruments and cancellation route. Comparing those, rather than headline prices, is what tells you which subscription actually fits how you watch.
8 min readChoosing a VR adult platform comes down to three structural questions rather than price: which projection format it ships, which player app it requires, and whether your headset class can handle the file sizes involved. Most Japanese VR output is 180-degree stereoscopic, delivered by download through a dedicated player.
9 min readFANZA is the adult-facing digital storefront operated by DMM, separated from the general DMM brand in 2018. FANZA sells Japanese adult video as individual purchases, time-limited rentals and monthly subscription tiers, all censored to meet Japanese law. FANZA's help centre states the service cannot be used from outside Japan (checked 2026-08-03).
9 min readLegal Japanese adult video comes from licensed platforms in five broad categories: large domestic storefronts, studio-owned subscription sites, distributor catalogues, general video-on-demand services with an adult tier, and VR-first platforms. Which ones you can actually use depends on payment method, region checks and your own jurisdiction.
10 min readSearch quality on adult video platforms cannot be ranked in the abstract, because each engine fails differently on ID formatting, romanised titles and performer aliases. Running a fixed set of 15 known IDs through every platform you use, and recording how each one fails, gives an answer specific to what you actually search for.
7 min readSubscribing to a Japanese adult platform from abroad depends on three independent gates: account registration, a payment instrument the platform will accept, and playback permitted from your IP address. A platform can pass two gates and still fail on the third, which is why overseas access varies so much between services.
9 min readBoth labels mark a release assembled from previously published material. BEST usually organises that material around one performer, while soshuhen (総集編) organises it around a series, theme or time period. Neither indicates new footage, and both reuse existing product numbers' content.
4 min readWhether 4K looks different from 1080p depends on screen size, viewing distance, encode quality and whether the source was genuinely captured at 4K. On a phone or a laptop at normal distance the gap is small to invisible; on a large screen up close, or in VR, it is obvious.
7 min readJapanese adult video genres divide along four independent axes: who the lead is, what the setting is, how the title was produced, and what the release format is. A single title carries a label from each axis, which is why genre lists look overwhelming until you separate the layers.
4 min readAttackers is a Japanese adult video label built on scripted dramatic fiction — suspense, confrontation and revenge narratives — rather than on the polished performer showcases of the flagship labels. Its codes run across several long-lived prefixes including ADN, SHKD, RBD and RBK.
5 min readPhones can play VR adult content in three ways: flat on the screen with panning, stereoscopically inside a slot-in viewer, or as a source streaming to a real headset. Only the viewer approaches what the content was shot for, and its optics and tracking impose hard limits.
7 min readChoosing a VR headset for video comes down to four properties: angular resolution, comfort across a long session, whether the device can play files locally or must stream from a computer, and how much friction stands between you and playback. Category matters more than model.
8 min readEnglish metadata for Japanese adult video is generated per catalogue with no shared standard, so the same tag, title or performer name arrives differently on every site. In our index, 197 of 2,333 named performers carry a romanised name that disagrees with the name recorded in another language's metadata for the same person.
5 min readConsent in Japanese adult video moved from informal agreement toward documented permission with a defined scope, verified identity and age, a withdrawal period around release, and separate approval for re-licensing. Enforcement is now split between statute, industry associations and the distribution platforms that carry the finished work.
7 min readDokusen (独占) means a platform is the only one carrying a title, senko (先行) means it has the title early, and mihodai (見放題) means the title is included in a subscription rather than sold separately. All three describe a commercial arrangement, and all three can change without the title changing.
4 min readE-BODY is a Japanese adult video label whose casting is organised explicitly around body type, with a focus on fuller-figured and bust-forward performers. Its main numbering line ran under the EBOD prefix from 2007 until 2023 and continues as EBWH, and its house style puts the physical premise ahead of narrative.
5 min readDistribution in Japanese adult video passed through four regimes — rental tape, retail disc, paid download, and streaming — and each shift moved control from the party that owned the shelf to the party that owned the catalogue. Production economics, catalogue size and discoverability changed with every handover.
6 min readAuto-translated subtitles are reliably good enough to follow who is doing what and why, and unreliable for wordplay, honorifics, names, and any line where a mistranslation changes the meaning. Quality is limited less by the translation stage than by the speech recognition feeding it.
6 min readAwards and rankings in Japanese adult video come from four different mechanisms — trade awards judged by juries or distributor votes, retailer charts driven by sales, platform charts driven by views, and fan votes driven by turnout. Each measures a different quantity, which is why they routinely disagree.
7 min readMosaic censorship replaces a region of each frame with a grid of solid blocks, each block painted the average colour of the pixels it covered. Averaging is a one-way operation applied to the master before encoding, so the detail inside a block is discarded at source and no later process can recover it.
7 min readDebuting in Japanese adult video runs through an agency rather than a studio: an applicant signs with a talent agency, the agency negotiates appearances with makers, and a maker builds a marketed launch around a first title. The published debut date marks that release, not necessarily a first appearance on camera.
6 min readProduction companies and talent agencies are separate businesses with separate incentives. The production company plans, finances and shoots a title and owns the resulting master; the agency represents the performer, negotiates her contracts and manages her schedule across every company she works with.
6 min readJapan's adult video business splits across five distinct roles: production companies that shoot, makers and labels that fund, own and code the titles, talent agencies that manage performers, review bodies that approve masters before release, and retail platforms that sell to viewers. Each holds different decision rights.
7 min readIdeaPocket is a major Japanese adult video label known for polished, star-led productions with a strong dramatic finish. Its main numbering line has moved through successive prefixes including IPZ, IPX and IPZZ, each one replacing the last rather than running alongside it.
5 min readMost Japanese adult video vocabulary is catalogue metadata, not description. Terms like hitozuma, shirouto, tanpin and dokusen mark a title's genre slot, production format or distribution status — this glossary defines the 30 that newcomers most often read as literal statements of fact.
7 min readkawaii* is a Japanese adult video label with a bright, idol-adjacent presentation — high-key lighting, colourful art direction and youthful styling in the pop-idol tradition. Its main numbering line moved from KAWD to CAWD and, in 2026, from CAWD to CAWB.
4 min readJapanese bust-size tags are marketing tiers, not measurements. Kyonyu covers the broad large range, bakunyu marks the top of it and hinnyu the small end, but no distributor publishes the cup size where one tier becomes the next, so the boundaries move between catalogues.
4 min readMadonna is a Japanese adult video label devoted to the married-woman (人妻, *hitozuma*) genre, casting adult women in domestic and relationship-driven dramas. Its main numbering line has moved through prefixes including JUY, JUL, JUQ and now JUR, each replacing the one before it.
5 min readSwitching between phone and desktop changes four things: how much of the picture your eyes can resolve, what the device can decode without stuttering, how much control you have over playback and subtitles, and what quality the site decides to send you. The content itself is identical.
7 min readMOODYZ is a major Japanese adult video label built around long-running numbered series covering an unusually broad range of genres. Its main numbering line has moved through prefixes including MIDD, MIDE, MIDV and now MIDA, with parallel prefixes such as MIAA and MIAB running as separate lines.
5 min readPRESTIGE is a Japanese adult video production company that operates as its own group rather than as one label inside a larger portfolio, and it runs a large number of series-specific code prefixes. Its best-known main line has moved through ABP, ABW and ABF.
5 min readJapanese adult video titles are built from a small set of recurring components: a series name, a subject label, one or two intensifier words, a numeric claim and a status marker. Learning roughly twenty of these words makes most titles parseable without translating them.
5 min readS1 NO.1 STYLE is a flagship Japanese adult video label known for glossy, high-budget productions built around performers under exclusive contract. Its main numbering line has run through a sequence of four-letter prefixes, including SNIS, SSNI, SSIS, SONE and now SNOS, each replacing the last.
5 min readSOD Create is a Japanese adult video production company known for variety-show and documentary-style formats rather than conventional scripted scenes. Its best-known numbering line now runs under the START prefix, which succeeded STARS and before that STAR, alongside several SD-prefixed series lines.
5 min readTanpin (単体) marks a release built around one billed performer, kikaku (企画) one built around a concept. Kikaku tanpin sits between them: a named performer fronts a concept-driven production. The distinction describes how a title was commissioned, not its quality.
4 min readJapan's adult video business is organised around production companies that each run several separate labels. The label, not the parent company, sets the house style, the series structure and the letter prefix that opens every product code, which is why one performer appears under many different codes.
6 min readVideo resolution is the size of the pixel grid that makes up each frame, written either as a vertical line count like 1080p or a horizontal one like 4K. Resolution sets a ceiling on how much detail a file can hold, but it never guarantees that the detail is actually there.
8 min readVR video files are described by three independent properties: the field of view they cover (180 or 360 degrees), whether they are stereoscopic and how the two eye images are packed into one frame, and the frame's pixel dimensions. Misreading any one of the three is why playback looks wrong.
9 min readGenerated content affects the adult video industry in three separate ways: fully synthetic material competes on cost with produced material, face-swapped material attaches a real person's likeness to work they never performed, and neither is reliably distinguishable from the outside — which turns provenance, not appearance, into the only usable evidence.
7 min readChijo (痴女) describes a female-led dynamic in which the woman drives the scene and directs what happens. As a catalogue tag it marks a reversal of the genre's default power arrangement, and it says nothing about the performer's age, appearance or marital framing.
4 min readHitozuma (人妻) literally means "another person's wife", but as a catalogue tag it marks a genre slot: an adult female lead in a domestic setting, styled as a settled married woman. It describes how a title was marketed, not the verified marital status of anyone in it.
4 min readJukujo (熟女) means "mature woman" and marks a marketing age band rather than a measured threshold. No distributor publishes the birthday at which a performer becomes jukujo, so the label reflects how a release was positioned, not a verified age.
4 min readShirouto (素人) means "layperson" or "non-professional", but as a catalogue tag it marks a production style rather than an employment status: handheld framing, unscripted delivery and a casting premise. Commercially distributed shirouto titles are still professionally produced, contracted and age-verified.
4 min readAn exclusive performer (専属, *senzoku*) is contracted to release through a single label for a fixed term, appearing on no rival label during it, in exchange for a guaranteed release slot, production budget and promotion. The arrangement shapes casting, cover billing and code prefixes across the whole industry.
6 min readNTR is an abbreviation of netorare (寝取られ), a Japanese narrative frame in which someone's partner is taken by another person, told from the viewpoint of the one who loses them. As a catalogue tag it marks a story structure and a point of view, not any particular act.
4 min readSubtitles for Japanese adult video come from a handful of structurally different production paths — distributor-supplied, caption-derived, human translation by ear, and machine transcription plus machine translation. The quality gap between two subtitle tracks is almost entirely explained by which path produced them.
6 min readChoosing a video player comes down to four properties rather than brand preference: codec and container support, hardware decoding on your specific device, subtitle handling, and VR projection support. Match those four to what you actually play, and the "best player" question answers itself.
6 min readBitrate is the amount of data used to describe each second of video, and it decides how much of the detail a resolution promised actually survives compression. Given the same file size, a higher-bitrate 1080p encode will usually look better than a bitrate-starved 4K one.
8 min readJapanese adult video is censored because Japan's Penal Code prohibits distributing obscene material without defining what obscenity looks like. The mosaic is the industry's own compliance answer, enforced before release by review bodies and by retailers who will not stock an unreviewed title, not by any statute specifying pixel sizes.
7 min readTravelling breaks online services because the signals that identify your country stop agreeing: your IP address moves while your account country, your card's issuing country and your app store storefront stay where they were. Each service reads a different signal, so failures appear one at a time rather than all at once.
6 min readAdult content is regulated through five distinct mechanisms — production rules, distribution and obscenity law, platform obligations, network-level access blocking, and payment restrictions — and countries combine them differently. Understanding which lever a country pulls explains far more than memorising individual statutes, which change frequently.
18 min readAge verification requirements oblige a platform to establish that a visitor is an adult by some means stronger than a self-declared checkbox. The methods range from document checks and facial age estimation to double-blind attestation tokens, and they differ enormously in how much identifying data leaves the user's control.
11 min readA DNS leak occurs when your device sends domain name lookups to a resolver you did not intend to use, usually your internet provider's, revealing every site you are about to visit. Checking means confirming which resolver actually answers your queries, at the operating system, browser and network level.
5 min readGeoblocking decides your country from several independent signals — the IP address mapped through a geolocation database, the CDN edge that received the request, the account's registered country, the card's issuing country and the app store storefront — and applies a rule before the origin server is ever reached. Those signals frequently disagree.
6 min readMalicious advertising reaches adult sites through automated ad exchanges that resell inventory several layers deep, so the site operator never reviews the creative that runs. Attackers buy that inventory, cloak the payload from reviewers, and target the visitor directly with fake players, forced redirects and notification hijacks.
5 min readMalware on streaming sites almost never arrives as an obvious program. It arrives as something the visitor is already expecting — a missing codec, a player update, a browser update, a CAPTCHA, a subtitle file or a download helper — because the disguise only has to survive the few seconds before the click.
5 min readCard safety on an adult platform depends on who handles the number, not on how the site looks. Sites that hand card entry to a payment processor never store the number themselves, and the larger exposure is usually the surrounding data — email address, billing name and purchase history — which the merchant does keep.
5 min readClearing browsing history removes one of roughly a dozen records your device keeps. Cache, cookies, site storage, service worker data, autofill entries, the DNS cache, synced copies on your account and operating system thumbnails all persist independently, and network-side logs are outside your control entirely.
5 min readFake subscription pages are identified by their domain, not their design. Read the registered domain immediately left of the top-level domain, check whether the card form is served by a payment processor rather than the page itself, and treat any deadline, threat or unexpected re-entry of card details as the primary warning sign.
5 min readVirtual cards are card numbers generated as aliases to a real account, usually with a spending limit and often locked to one merchant. They contain the damage from a leak and give you a hard stop on recurring billing, but they do not make a purchase anonymous and do not cancel a subscription.
5 min readIncognito mode hides your activity from other people who use the same device, by discarding history, cookies and form data when the session ends. It hides nothing from your internet provider, your employer's network, or the websites themselves, all of which see the same traffic as in a normal window.
5 min readCoverage of the starlikeness index is narrow and measurable: 241,792 face vectors, but only 2,333 identity records, of which 1,707 carry a real profile. Roughly 1% of indexed faces are linked to a name, and the index measures aggregator exposure rather than industry output.
6 min readSimilar-performer lists in the starlikeness index contain 7,004 references that point to identities with no name attached — 52% of all such references. They appear in 1,899 performers' lists and resolve to 589 distinct unnamed identities, which is what an incomplete industry record looks like from inside.
6 min readA similarity score is the cosine similarity between two 512-number face vectors, running from 0 to 1. On our data, 0.40 is the threshold above which two faces are treated as probably the same person. Below it, a result is simply the nearest available stranger.
6 min readData in the starlikeness index comes from cover images and video stills published on 106 public sites, plus performer metadata parsed from public profiles. Representative images come from only four hosts, one of which supplies 2,246 of 2,333 records, so coverage reflects aggregator exposure.
6 min read