Nishi Family › Compare › Performer Identity Search -- text plus media, is it her, only her

Nishi Compare · capabilities, evidence and delivery

Performer Identity Search -- text plus media, is it her, only her

Compare capabilities and delivery evidence. Source discovery, qualification, deployment and accepted outcomes are separate; an open contract names work still to verify.

The roadmap-binding board behind /compare/performersearch. Nishi cells are MEASURED against organs on disk; rival cells are today's live probes or documented capability.

Overview

The purpose, declared position and evidence coverage of this domain. Source presence and completed acceptance are different measures.

Where we are. MEASURED 2026-09-02, not recalled. The Nishi text index returns 0 relevant of 30 for both julia kyoka and chloe lamb. The entity organs exist and are unwired to any performer record: nx_entity_dossier (card, portfolio, coverage), nx_entity_seeds (one page to the other pages about the entity), nx_entity_media (multi-image gather with a javdatabase adapter), nx_page_ingest (entity-depth ingester). Reverse image is a COPY detector (nx_imgsearch, 992 permil over 17 transforms) whose identity tier is declared in nx_semruler_lib (sr_hit_is_identity) and published UNMEASURED at permil -1 because no image encoder exists; nx_face_count is a cardinality detector only. Clean-serve delivers a KVS page's native player 5 of 5 and plays in-browser 1 of 5 until the CS7 relay. The referees, live today: Bing Videos 54 records for chloe lamb (51 titles with both names, 48 adult hosts, 20 re-hosted preview clips) and 18 for julia kyoka (4 with both names, the rest by alias Julia Boin and JAV codes); Yandex Video about 17 each; Yandex Images CBIR face search documented; Google Videos a JavaScript shell to the probe.

Where we need to go. Know the entity as data (names, aliases, catalogue codes, documented sources), grow its corpus from every query through the organs that already exist, rank the sovereign index by entity, hold a documented reference set, and answer is-it-her on an image, a gif and a video with a sovereign encoder and a referee-calibrated verdict, so results are filtered to the performer and delivered through clean-serve with no attack vectors. Every referee number is a beat on a time spine, never a recalled claim.

The unit. One unit is one seat-session of build plus the gate that proves it. Estimates are units, never hours.

Research bar. Bing Videos and Visual Search is measured on video records returned for a performer name with SafeSearch off, titles carrying both names, adult-host share, re-hosted preview clips; reverse image by upload [@bing-vs]. Theirs: chloe lamb 54 records, 51 titles with both names, 48 adult hosts, 20 preview clips; julia kyoka 18 records, 4 with both names, the rest by alias and catalogue code (measured 2026-09-02). Ours: UNPINNED -- a hosted search engine publishes no npm package and no GitHub release, so there is no version to bank; the bar is the live measurement above, re-read by the P8 referee beat. Ours: 0 relevant of 30 from the Nishi index for both names; reverse image copy tier 992 permil over 17 transforms, no face tier.

Research bar. Yandex Images CBIR is measured on content-based image retrieval including same-face lookup, plus Yandex Video results per name [@yandex-images]. Theirs: CBIR face lookup documented and present on the page; Yandex Video about 17 items per name (measured 2026-09-02). Ours: identity tier declared in nx_semruler_lib and published UNMEASURED at permil -1: no encoder.

Research bar. Google Lens and Google Videos is measured on reverse image, lookalike retrieval, knowledge-panel disambiguation, video results per name [@google-lens]. Theirs: documented; the video results page answers a JavaScript shell to a non-browser client, so counts are UNMEASURED from the NAS until the Waterfox referee beat runs. Ours: no entity record, no disambiguation signal; the dossier organ picked a Wikipedia page about stone idols for chloe lamb.

Latest recorded release

No valid dated release entry is recorded for this domain.

Release entries describe recorded changes; they do not establish that every capability passed evaluation.

8 of 18 capability source symbols found|2 of them authored stronger-claims|10 open|coverage 444/1000|adoption 3 full / 5 partial

Evidence profile — what the gaps on this board actually are

Measured by nx_swcompare_evidence, read back by nx_evprofile_lib. Every figure is a count with its denominator — there is deliberately no score, no grade and no percentage anywhere in this band, because a stored scalar is a field a seat can edit and a counted partition is not.

proven able to fail counts the gates that have a RECORDED RED — nx_gate_bite mutated the gate subject, rebuilt it, watched the gate go red, and that record is inside the shared TTL. never bitten is its complement over the same denominator: those gates ran and were green, and nothing has ever shown them able to detect anything, so their green is a statement about this run and not about the gate. green at 0/0 is a separate and much weaker observation — the gate printed GREEN on a zero denominator, so its own tooth counter says it examined nothing. A gate can be green, non-zero, and still never bitten; that is the common case and it is now visible instead of implied.

partition: grounded + unsupported = 8 vs present 8 · named + unnamed + withheld = 10 vs open 10 · both reconcile

liar-kill conj=GPqN · first failing conjunct: quorum

graded document: BUILDROOT tree, 5773 bytes · gates map: PRIMARY · stamped 35d 6h ago · STALE against the shared TTL of 604800s — this is evidence about a build nobody can vouch for now · source ../knowledge/status/evstamp_performersearch.verdict

Gap class	What it is, and the work it names
CLAIM-ONLY	No declared gate ran for this board, so every capability mark here rests on a symbol existing in source and on nothing executing. The remedy is to WRITE a `.gates` map naming gates that already run — authoring, not research.
STALE	This profile is older than the shared TTL, or its clock ran backwards. It is reported as unusable rather than as fresh, because being wrong in the direction of refusing is the safe direction.

This band reports the referee counts. The per-axis worklist rows — which axis is unsupported, which watch is ready to flip, which gap is unnamed — are printed by nx_swcompare_evidence performersearch itself and are not carried on the stamp, so this page names the classes and the producer names the rows. That split is stated rather than hidden: a count without a worklist is not actionable, and this band is honest about which half of that it is.

Production map

Follow the dependencies, declared acceptance criteria and recorded priorities. Inspect source binding before treating a rank as executable work.

Ranking source binding: PLAN_MATRIX_BOUND_ONLY. Recorded priorities require current acceptance evidence and resource checks before execution.

Ranking matches the captured plan and matrix only. Latest execution outcome, research freshness, accepted delivery and investment return are unverified.

Recorded priority estimates

Order from nx_compare_rank (nx_dr_ocm: (deficit + cost-of-delay + option + enables) x sponsor x self-sufficiency x momentum / cost). FINISH rows are rungs whose symbol is present but whose organ is short of full adoption: listed before new work by this heuristic. Priority is not measured delivery cost or execution readiness. Stamp: # asof=1790081606 domain=performersearch target_version=v0.1 entity rungs=10 done=0 open=10 finish=0 ranker=nx_dr_ocm

#	Stage	Rung	Priority	Derivation
#1	v0.1 entity	Performer entity record and alias table as data (P0) _ABSENT_:pe_load dossier	5000	v=50 m=1 c=10
#2	v0.1 entity	Seed harvest through the existing organs (P1) _ABSENT_:pe_seed_harvest dossier	4000	v=60 m=1 c=15
#3	v0.1 entity	Documented reference set with provenance (P3) _ABSENT_:pe_refset dossier	3000	v=30 m=1 c=10
#4	v0.1 entity	Entity-aware text ranking (P2) _ABSENT_:pe_query_expand dossier	200	v=3 m=1 c=15
#5	later	Sovereign face and image encoder (P4) _ABSENT_:fe_embed dossier	3000	v=90 m=1 c=30
#6	later	Identity verdict with a referee panel (P5) _ABSENT_:pe_identity_verdict dossier	2000	v=40 m=1 c=20
#7	later	Results filtered to the performer (P7) _ABSENT_:pe_filter dossier	1000	v=10 m=1 c=10
#8	later	Referee census beats (P8) _ABSENT_:pe_referee_census dossier	100	v=1 m=1 c=10
#9	later	Clean-serve as the delivery sandbox (P9) _ABSENT_:pe_clean_delivery dossier	100	v=1 m=1 c=10
#10	later	Video and gif identity by sampled frames (P6) _ABSENT_:pe_video_identity dossier	50	v=1 m=1 c=20

Declared roadmap — contract, acceptance, executor, effort

Rung	Closes with	Definition of done (pre-declared)	Executor	Est.
Performer entity record and alias table as data (P0)	_ABSENT_:pe_load	Pre-declared: knowledge/performers/<slug>.conf rows (canonical name, aliases, catalogue codes, documented source URLs, negative aliases that are OTHER people) read by ONE lib ruler; julia kyoka carries the alias Julia Boin and the JAV codes Bing resolves through, chloe lamb carries the artist and the footballer as negatives. Gate: a record loads, a missing record refuses by name, a negative alias is never matched as the entity. Organ: runtime/nx_performer_lib.nx	claude	1 u
Seed harvest through the existing organs (P1) after P0	_ABSENT_:pe_seed_harvest	Pre-declared: for a record, nx_entity_seeds over each documented source, nx_page_ingest at bounded depth, nx_entity_media for the images (javdatabase and wikimedia adapters exist), all under the mediaingest pacer; every discovered URL lands in the performer's corpus with its source and class, and the clean-serve corpus (knowledge/cleanserve_sites.conf) is a view of it. Gate: a fixture record with two sources yields the planted seeds and no duplicate. Organ: runtime/nx_performer_harvest.nx	claude	1.5 u
Entity-aware text ranking (P2) after P0	_ABSENT_:pe_query_expand	Pre-declared: the query is expanded with the record's aliases and codes, negatives are demoted, and host classes (tube and database hosts) are weighted; measured against Bing's baseline on the same two names (54 and 18 records): the relevant-of-30 count must move from 0 to a published number, and a NEGATIVE CONTROL query for an unrelated common name must not gain adult hosts. Organ: runtime/nx_performer_rank.nx	claude	1.5 u
Documented reference set with provenance (P3) after P1	_ABSENT_:pe_refset	Pre-declared: per performer, N reference images and poster frames with source URL, fetch date and licence tier stored under CID; a set below a floor declared in conf refuses to be used for verdicts; javdatabase idol images are the first source for julia kyoka. Organ: runtime/nx_performer_refset.nx	claude	1 u
Sovereign face and image encoder (P4)	_ABSENT_:fe_embed	Pre-declared: a face detector plus an embedding in NishiLang on the no-float seat (the CLIP-class encoder slot media R10 msm_search and mediaingest R17 share), with a known-answer test against a banked reference and a published pair-accuracy on a labeled same-person and different-person set; below the pre-declared bar the lane stays unwired. Organ: runtime/nx_face_embed.nx	Local model	3 u
Identity verdict with a referee panel (P5) after P3,P4	_ABSENT_:pe_identity_verdict	Pre-declared: cosine against the reference set with a threshold calibrated on the labeled set, three outcomes (HER, NOT-HER, UNMEASURED), and a VLM judge panel as the independent referee per the consensus law; the verdict never publishes without the reference-set floor. Organ: runtime/nx_performer_identity.nx	claude	2 u
Video and gif identity by sampled frames (P6) after P5	_ABSENT_:pe_video_identity	Pre-declared: keyframes through the first-IDR decoder (mediaingest R14 vph_scene shares the sampler) and gif frames run through P5; a per-video verdict with the frame partition published (her, not-her, no-face); a fixture video of the performer and one of another person separate. Organ: runtime/nx_performer_identity.nx	claude	2 u
Results filtered to the performer (P7) after P5	_ABSENT_:pe_filter	Pre-declared: a result row is served only when its identity verdict is HER, or is served flagged when UNMEASURED; NOT-HER rows are dropped and counted; the count is published beside the query. Organ: runtime/nx_performer_rank.nx	claude	1 u
Referee census beats (P8) after P0	_ABSENT_:pe_referee_census	Pre-declared: per performer, a beat records Bing Videos, Bing Images, Yandex Video and Yandex Images counts, host partition, alias coverage and preview-clip counts on a time spine, and the Google row comes from the laptop Waterfox referee because Google serves a JavaScript shell to a non-browser; the sota rows above are re-graded from the spine, never by hand. Organ: runtime/nx_performer_referee.nx	claude	1 u
Clean-serve as the delivery sandbox (P9) after P7	_ABSENT_:pe_clean_delivery	Pre-declared: every result row is served through /clean with the CS7 relay and the real-browser referee measures playback; a result that cannot be delivered clean is a named class on the board, never a link-out. Organ: runtime/nx_performer_rank.nx	claude	1 u

Milestones

Milestone	Rungs	Cumulative
Know the entity · record, harvest, ranking, reference set	P0,P1,P2,P3	5 u
Is it her · encoder, verdict, video and gif, filtering	P4,P5,P6,P7	8 u
Referees and delivery · census beats and clean delivery	P8,P9	2 u

Ladder verdict. NOT DECLARED. This board names no dated best-in-class or frontier target and no rung roles (sotatarget and rungrole rows on its plan); the ranker labels it NO-LADDER until it does, and until then its rungs climb toward a target nobody has written down. Bars read NO-BAR against the current month 2026-10.

Bar freshness, counted
Class	Count	Note
bars	0	sotabar rows on the plan
fresh	0	the bar month is the current month
attested	0	an older bar with a this-month barscan reporting no newer entry
stale	0	a this-month scan saw a newer month the bar ignores
unattested	0	nobody looked this month
malformed	0	a sotabar row missing a field
duplicate	0	bar ids an earlier row already carries
bars, summed	0	RECONCILED: the parts sum to the total 0
current month	2026-10	the month every bar is judged against

Treatment verdict. NOT DECLARED. No rung on this board declares its three-step treatment (a treat row naming the MCP row, the API route, the primitive it composes, the workflow that composes it and the agent beat that runs it); the ranker labels every rung TREATMENT-UNDECLARED until it does.

Inspect a rung and its prerequisites

Declared nodes 10. Rank input binding: PLAN_MATRIX_BOUND_ONLY. Dependency order is authored. Implementation, acceptance evidence, authority and resource readiness are unverified. No action is recommended or dispatched here.

Plan SHA-256 25c405eedade33f6820e033cf364775b304ee3f430add9e99d8a71bea725ad44. Target rows 0; role rows 0. Existing risks and release worklog retain their own scope; no node completion is inferred.

Runtime observation coverage: 0 of 10 rungs have current, identity-bound execution observations. Observation may report RED. Accepted completion remains unknown; source presence, qualification and deployment do not establish it.

Use Enter or Space on a rung to inspect its contract. Prerequisite links locate another rung in this list; open its summary to inspect it. Estimates are authored effort, not forecasts.

P0 — Performer entity record and alias table as data
Prerequisites: None declared; this does not establish execution eligibility.
Contract: _ABSENT_:pe_load
Acceptance: Pre-declared: knowledge/performers/<slug>.conf rows (canonical name, aliases, catalogue codes, documented source URLs, negative aliases that are OTHER people) read by ONE lib ruler; julia kyoka carries the alias Julia Boin and the JAV codes Bing resolves through, chloe lamb carries the artist and the footballer as negatives. Gate: a record loads, a missing record refuses by name, a negative alias is never matched as the entity. Organ: runtime/nx_performer_lib.nx
Authored effort: 1. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P1 — Seed harvest through the existing organs
Prerequisites: P0 (acceptance unverified)
Contract: _ABSENT_:pe_seed_harvest
Acceptance: Pre-declared: for a record, nx_entity_seeds over each documented source, nx_page_ingest at bounded depth, nx_entity_media for the images (javdatabase and wikimedia adapters exist), all under the mediaingest pacer; every discovered URL lands in the performer's corpus with its source and class, and the clean-serve corpus (knowledge/cleanserve_sites.conf) is a view of it. Gate: a fixture record with two sources yields the planted seeds and no duplicate. Organ: runtime/nx_performer_harvest.nx
Authored effort: 1.5. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P2 — Entity-aware text ranking
Prerequisites: P0 (acceptance unverified)
Contract: _ABSENT_:pe_query_expand
Acceptance: Pre-declared: the query is expanded with the record's aliases and codes, negatives are demoted, and host classes (tube and database hosts) are weighted; measured against Bing's baseline on the same two names (54 and 18 records): the relevant-of-30 count must move from 0 to a published number, and a NEGATIVE CONTROL query for an unrelated common name must not gain adult hosts. Organ: runtime/nx_performer_rank.nx
Authored effort: 1.5. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P3 — Documented reference set with provenance
Prerequisites: P1 (acceptance unverified)
Contract: _ABSENT_:pe_refset
Acceptance: Pre-declared: per performer, N reference images and poster frames with source URL, fetch date and licence tier stored under CID; a set below a floor declared in conf refuses to be used for verdicts; javdatabase idol images are the first source for julia kyoka. Organ: runtime/nx_performer_refset.nx
Authored effort: 1. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P4 — Sovereign face and image encoder
Prerequisites: None declared; this does not establish execution eligibility.
Contract: _ABSENT_:fe_embed
Acceptance: Pre-declared: a face detector plus an embedding in NishiLang on the no-float seat (the CLIP-class encoder slot media R10 msm_search and mediaingest R17 share), with a known-answer test against a banked reference and a published pair-accuracy on a labeled same-person and different-person set; below the pre-declared bar the lane stays unwired. Organ: runtime/nx_face_embed.nx
Authored effort: 3. Executor kind: Local model. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P5 — Identity verdict with a referee panel
Prerequisites: P3 (acceptance unverified), P4 (acceptance unverified)
Contract: _ABSENT_:pe_identity_verdict
Acceptance: Pre-declared: cosine against the reference set with a threshold calibrated on the labeled set, three outcomes (HER, NOT-HER, UNMEASURED), and a VLM judge panel as the independent referee per the consensus law; the verdict never publishes without the reference-set floor. Organ: runtime/nx_performer_identity.nx
Authored effort: 2. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P6 — Video and gif identity by sampled frames
Prerequisites: P5 (acceptance unverified)
Contract: _ABSENT_:pe_video_identity
Acceptance: Pre-declared: keyframes through the first-IDR decoder (mediaingest R14 vph_scene shares the sampler) and gif frames run through P5; a per-video verdict with the frame partition published (her, not-her, no-face); a fixture video of the performer and one of another person separate. Organ: runtime/nx_performer_identity.nx
Authored effort: 2. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P7 — Results filtered to the performer
Prerequisites: P5 (acceptance unverified)
Contract: _ABSENT_:pe_filter
Acceptance: Pre-declared: a result row is served only when its identity verdict is HER, or is served flagged when UNMEASURED; NOT-HER rows are dropped and counted; the count is published beside the query. Organ: runtime/nx_performer_rank.nx
Authored effort: 1. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P8 — Referee census beats
Prerequisites: P0 (acceptance unverified)
Contract: _ABSENT_:pe_referee_census
Acceptance: Pre-declared: per performer, a beat records Bing Videos, Bing Images, Yandex Video and Yandex Images counts, host partition, alias coverage and preview-clip counts on a time spine, and the Google row comes from the laptop Waterfox referee because Google serves a JavaScript shell to a non-browser; the sota rows above are re-graded from the spine, never by hand. Organ: runtime/nx_performer_referee.nx
Authored effort: 1. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.
P9 — Clean-serve as the delivery sandbox
Prerequisites: P7 (acceptance unverified)
Contract: _ABSENT_:pe_clean_delivery
Acceptance: Pre-declared: every result row is served through /clean with the CS7 relay and the real-browser referee measures playback; a result that cannot be delivered clean is a named class on the board, never a link-out. Organ: runtime/nx_performer_rank.nx
Authored effort: 1. Executor kind: claude. Responsible, accountable and verifier not established. Inspect retained worklog.
Runtime observation: UNKNOWN; outcome UNKNOWN; observed epoch -1. Declared owner: . Receipt , SHA-256 . Identity-bound observation is not accepted completion. Qualification and deployment remain unverified by this projection.

Learning and practice paths

No structured learning path is declared for this plan. Existing research, roadmap and worklog remain available above.

Rung dossiers — every record that bears on a rung, in one place

Joined by symbol, never typed. Each dossier is assembled on every publish from the rows this board already carries: the plan's rung row, the matrix row that carries its contract symbol (its status re-measured by the one symbol ruler), the gate the organ declares and the verdict the evidence referee recorded on its last run, the reference marks in the rung text, the design rows that name the rung and the work log filed against it. The capability matrix links each cell to its dossier and each dossier links back to its cell and forward to the rungs it stands on; a dependency naming no rung on this plan is written as UNRESOLVED rather than linked. Absences are named: NO-MATRIX-ROW, NO-GATE-DECLARED, UNRECORDED.

Gate verdicts: UNSTAMPED — no per-gate verdict artifact is published for this domain yet (nx_swcompare_evidence performersearch writes it beside its stamp); every gate below reads UNRECORDED until it runs.

Gate declarations: knowledge/organ_gate.conf read from this working directory (1 rows) and from the estate root (196 rows); rows in both counted once: 0; declarations read: 197. Every declaration in either copy counts, so a stub beside the publisher can never hide the root.

P0 Performer entity record and alias table as data · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_load` · executor claude · 1 u · stands on nothing
Definition of done	Pre-declared: knowledge/performers/<slug>.conf rows (canonical name, aliases, catalogue codes, documented source URLs, negative aliases that are OTHER people) read by ONE lib ruler; julia kyoka carries the alias Julia Boin and the JAV codes Bing resolves through, chloe lamb carries the artist and the footballer as negatives. Gate: a record loads, a missing record refuses by name, a negative alias is never matched as the entity. Organ: runtime/nx_performer_lib.nx
Matrix cell	Performer entity record and alias table · `runtime/nx_performer_lib.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_lib`; declare `nx_performer_lib TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	2026-09-02 measure BASELINE 2026-09-02 (this session): Nishi web search returned 0 relevant of 30 for julia kyoka (Penguin, Wikipedia Julia Scott, a German TV page, Lewd Swedes Julia Crown) and 0 of 30 for chloe lamb (Lemmy, an obituary, NASA, a handbag blog). Bing Videos: chloe lamb 54 records, 51 titles carrying both names, 48 on adult hosts (camwhores 13, tnaflix 12, peekvids 7, txxx 4, nuvid 4), 20 with a Bing-re-hosted preview clip; julia kyoka 18 records, 4 titles with both names, the rest by the alias Julia Boin and JAV codes REBD-425 and PPPD-335. Bing Images 34 and 100 tiles with Visual Search. Yandex Video about 17 each. Yandex Images CBIR present. Google Videos: a JavaScript shell to the probe. DDG 10 results. Estate incumbents: nx_entity_dossier, nx_entity_seeds, nx_entity_media (javdatabase adapter), nx_page_ingest exist and are unwired to any performer record; nx_imgsearch copy tier 992 permil; nx_semruler_lib sr_hit_is_identity declared, permil -1; nx_face_count cardinality only; no encoder. 2026-09-02 queue OPERATOR DIRECTION 2026-09-02: integrate clean-serve with the media capabilities so a known performer is searched by text and media, the corpus grows, the estate knows if it is her in video, image or gif like Bing does, results are only hers, and the engines are measured on /compare. Test entity: Julia Kyoka first (documented sources, javdatabase adapter, catalogue codes), Chloe Lamb second (the live clean-serve corpus). First-byte blockers named as debt rows D1 to D7 above. 2026-09-02 measure INCUMBENT MEASURED: nx_entity_dossier chloe lamb returned found=false and picked en.wikipedia.org Celtic_stone_idols as the landing page (card fields 0, authority 0, coverage permil 0 over 19,811 indexed matches) -- the dossier organ is real, runs, and has no performer record to anchor on; the julia kyoka run is in flight. The referee mirrors for the bar rows below were fetched through nx_research_fetch and pinned by hash. 2026-09-02 measure nx_entity_dossier julia kyoka: found=true on www.javdatabase.com (pick kind 2, card fields 8: JP name JULIA, DOB 1987-05-25, Tokyo, measurements, cup, height, hair; authority_total 1051 titles; codes sampled STOL-139, NHDTA-633) -- the estate CAN anchor a database-documented performer today, which is why Julia Kyoka is the first test entity and the P0 record starts from this card plus the alias Julia Boin the engines resolve through. Two defects measured in the same run: coverage.permil reads 18716 (indexed_matched 19671 over authority 1051), a ratio above 1000 that cannot be a fraction, because indexed_matched counts raw name hits across the index rather than entity-linked pages; and for chloe lamb the organ picked a Wikipedia page about Celtic stone idols with found=false, so its landing choice has no negative-alias guard. Both feed P0 and P2.

P1 Seed harvest through the existing organs · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_seed_harvest` · executor claude · 1.5 u · stands on P0
Definition of done	Pre-declared: for a record, nx_entity_seeds over each documented source, nx_page_ingest at bounded depth, nx_entity_media for the images (javdatabase and wikimedia adapters exist), all under the mediaingest pacer; every discovered URL lands in the performer's corpus with its source and class, and the clean-serve corpus (knowledge/cleanserve_sites.conf) is a view of it. Gate: a fixture record with two sources yields the planted seeds and no duplicate. Organ: runtime/nx_performer_harvest.nx
Matrix cell	Seed harvest through the existing organs · `runtime/nx_performer_harvest.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_harvest`; declare `nx_performer_harvest TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P2 Entity-aware text ranking · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_query_expand` · executor claude · 1.5 u · stands on P0
Definition of done	Pre-declared: the query is expanded with the record's aliases and codes, negatives are demoted, and host classes (tube and database hosts) are weighted; measured against Bing's baseline on the same two names (54 and 18 records): the relevant-of-30 count must move from 0 to a published number, and a NEGATIVE CONTROL query for an unrelated common name must not gain adult hosts. Organ: runtime/nx_performer_rank.nx
Matrix cell	Entity-aware text ranking · `runtime/nx_performer_rank.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_rank`; declare `nx_performer_rank TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P3 Documented reference set with provenance · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_refset` · executor claude · 1 u · stands on P1
Definition of done	Pre-declared: per performer, N reference images and poster frames with source URL, fetch date and licence tier stored under CID; a set below a floor declared in conf refuses to be used for verdicts; javdatabase idol images are the first source for julia kyoka. Organ: runtime/nx_performer_refset.nx
Matrix cell	Documented reference set with provenance · `runtime/nx_performer_refset.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_refset`; declare `nx_performer_refset TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P4 Sovereign face and image encoder · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:fe_embed` · executor Local model · 3 u · stands on nothing
Definition of done	Pre-declared: a face detector plus an embedding in NishiLang on the no-float seat (the CLIP-class encoder slot media R10 msm_search and mediaingest R17 share), with a known-answer test against a banked reference and a published pair-accuracy on a labeled same-person and different-person set; below the pre-declared bar the lane stays unwired. Organ: runtime/nx_face_embed.nx
Matrix cell	Sovereign face and image encoder · `runtime/nx_face_embed.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_face_embed`; declare `nx_face_embed TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P5 Identity verdict with a referee panel · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_identity_verdict` · executor claude · 2 u · stands on P3, P4
Definition of done	Pre-declared: cosine against the reference set with a threshold calibrated on the labeled set, three outcomes (HER, NOT-HER, UNMEASURED), and a VLM judge panel as the independent referee per the consensus law; the verdict never publishes without the reference-set floor. Organ: runtime/nx_performer_identity.nx
Matrix cell	Identity verdict with a referee panel · `runtime/nx_performer_identity.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_identity`; declare `nx_performer_identity TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P6 Video and gif identity by sampled frames · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_video_identity` · executor claude · 2 u · stands on P5
Definition of done	Pre-declared: keyframes through the first-IDR decoder (mediaingest R14 vph_scene shares the sampler) and gif frames run through P5; a per-video verdict with the frame partition published (her, not-her, no-face); a fixture video of the performer and one of another person separate. Organ: runtime/nx_performer_identity.nx
Matrix cell	Video and gif identity by sampled frames · `runtime/nx_performer_identity.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_identity`; declare `nx_performer_identity TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P7 Results filtered to the performer · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_filter` · executor claude · 1 u · stands on P5
Definition of done	Pre-declared: a result row is served only when its identity verdict is HER, or is served flagged when UNMEASURED; NOT-HER rows are dropped and counted; the count is published beside the query. Organ: runtime/nx_performer_rank.nx
Matrix cell	Results filtered to the performer · `runtime/nx_performer_rank.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_rank`; declare `nx_performer_rank TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P8 Referee census beats · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_referee_census` · executor claude · 1 u · stands on P0
Definition of done	Pre-declared: per performer, a beat records Bing Videos, Bing Images, Yandex Video and Yandex Images counts, host partition, alias coverage and preview-clip counts on a time spine, and the Google row comes from the laptop Waterfox referee because Google serves a JavaScript shell to a non-browser; the sota rows above are re-graded from the spine, never by hand. Organ: runtime/nx_performer_referee.nx
Matrix cell	Referee census beats · `runtime/nx_performer_referee.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_referee`; declare `nx_performer_referee TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

P9 Clean-serve as the delivery sandbox · cell WATCHING · gate NO-GATE-DECLARED

Contract	`_ABSENT_:pe_clean_delivery` · executor claude · 1 u · stands on P7
Definition of done	Pre-declared: every result row is served through /clean with the CS7 relay and the real-browser referee measures playback; a result that cannot be delivered clean is a named class on the board, never a link-out. Organ: runtime/nx_performer_rank.nx
Matrix cell	Clean-serve as the delivery sandbox · `runtime/nx_performer_rank.nx` · WATCHING decl
Gate	NO-GATE-DECLARED — `knowledge/organ_gate.conf` carries no row for `nx_performer_rank`; declare `nx_performer_rank TAB <gate>` and the ship loop proves it and the verdict lands here.
Work log	no log row names this rung yet

dossiers=10 cells_joined=10 no_matrix_row=0 (sum 10) gates_recorded=0 gates_unrecorded=0 no_gate_declared=10 (sum 10 of cells_joined) evgates=UNSTAMPED

Capability comparisons

Compare the field, search individual capabilities and open their source and adoption evidence. Documented presence does not establish comparative quality.

Position map — centrality and distinctiveness

The four-quadrant map the field uses for brand strategy (Dawar and Bagga, HBR June 2015), re-derived from this matrix on every publish. Centrality is the share of the category's feature mass a player covers, each feature weighted by how many hold it; distinctiveness is the average lead over each rival on the rows the player holds; breadth is the depth-weighted share of the whole matrix (the bubble); depth is how deeply the rows held are held; momentum is the day-over-day move off the spine (green rising, red falling, grey until day two); the dashed path runs first day → previous day → today. Dividers are the category means. Axes are fitted to the field of play, so read the tick numerals, not the frame. Rival marks are documented presence, so a rival's position reads the record, never its quality. The picture grades its own readability below; the table beside it is the same data for a screen reader or a second method.

Rows. 18 feature rows mapped of 18 matrix rows; 0 substrate rows (kernels, bounds and instruments the plan's rungrole marks substrate) are the estate's scaffolding, not category features, and stay off the map

Views. 2D cut: centrality, distinctiveness · 3D cube: centrality, distinctiveness, breadth · axes registered: centrality, distinctiveness, breadth, depth, momentum · a board picks its own in knowledge/compare/performersearch.cdmap (cut|x|y, cube|x|y|z); absent = the HBR defaults

bubble area = breadth · ring = momentum (green rising, red falling, grey until day two) · dashed = category means · axes fitted to the field of play: centrality 300–800, distinctiveness 200–700 of 0–1000 permil (the full range put every player in one corner)
readability of the 2D cut, self-graded by the layout ruler: label overlaps 0 · labels over marks 0 · off-canvas 0 · unresolved labels 0 · mark overlaps 0 (a fact of the data: two players that close are that close) · data spread 577 permil of the plot · quadrant words unseated 0
text contrast, measured with wcag2-ratio (floors from contrast.conf), light theme: labels 16.24 (floor 4.50) · notes 16.24 (floor 4.50) · quadrant words 5.89 (floor 4.50) · tick numerals 5.89 (floor 4.50) · axis titles 5.59 (floor 4.50) · dark theme: labels 13.78 (floor 4.50) · quadrant words 5.55 (floor 4.50) · axis titles 5.97 (floor 4.50) · dark classes under their floor 0 (one figure serves both themes: a dark shortfall is a token to fix, never a class to hide) · classes refused under their floor 0 (a refused class is not drawn; the scale classes are measured, never hidden) · export: SVG PNG (receipt, rendered by the estate's own rasteriser from this page)

axes fitted: centrality 300–800 · distinctiveness 200–700 · breadth 200–450 permil · farther bubbles are painted first, nearer ones over them · the floor shadow is each bubble's (x, y) at height 0
readability of the 3D cube, self-graded by the layout ruler: label overlaps 0 · labels over marks 0 · off-canvas 0 · unresolved labels 0 · mark overlaps 1 (a fact of the data: two players that close are that close) · data spread 393 permil of the plot

Nishi Bing-Videos Yandex-Images Google-Lens Adult-DBs
10 panels over 5 registered axes, every pair once (the lower axis on x, the higher on y) · each panel fitted to its own field of play, first and last tick numerals shown · no labels in a small cut, the legend names the colours; the grade below reports the mark terms only (MM summed over the panels, spread averaged)
readability of the small multiples, self-graded by the layout ruler: label overlaps 0 · labels over marks 0 · off-canvas 0 · unresolved labels 0 · mark overlaps 7 (a fact of the data: two players that close are that close) · data spread 497 permil of the plot

Player	Quadrant	centrality	distinctiveness	breadth	depth	momentum	Rows held	Days on spine	First seen
Nishi	Mainstream	631	427	333	750	0 since 2026-09-23	8	10	2026-09-15
Bing-Videos	Mainstream	657	427	333	750	0 since 2026-09-23	8	10	2026-09-15
Yandex-Images	Unconventional	394	600	259	933	0 since 2026-09-23	5	10	2026-09-15
Google-Lens	Mainstream	684	312	296	666	0 since 2026-09-23	8	10	2026-09-15
Adult-DBs	Aspirational	684	555	425	851	0 since 2026-09-23	9	10	2026-09-15

Day	Matrix rows reviewed	Players recorded
2026-09-15	18	5
2026-09-16	18	5
2026-09-17	18	5
2026-09-18	18	5
2026-09-19	18	5
2026-09-20	18	5
2026-09-21	18	5
2026-09-22	18	5
2026-09-23	18	5
2026-10-08	18	5

How to read this evidence. A Nishi source mark records whether the named symbol is declared in the selected source file. It does not establish working behavior or accepted delivery. A watching tag names an expected source entry; discovering it updates source coverage only. A dark tag means the file exists but the named symbol was not found: inspect the source and contract before assuming the capability is missing or renaming code. Peer codes are authored assessments; unassessed means no comparison is established, not absence. Qualification, deployment identity, runtime observations and acceptance must carry their own evidence. Adoption records the catalog's observed path through source, build, promotion, registration and invocation; libraries use importer evidence and gates use declared execution surfaces. A full catalog path does not establish quality, freshness or a successful user outcome. Partial and unknown states retain their stated next action. Census stamps: importers asof 1790637268, gate census asof 1790104131 (unix seconds; -1 = census absent).

Capability matrix — measured against source

◉ authored stronger claim; source present● present◐ partial○ absent? unassessed; excluded from relative comparison · click any capability for its evidence

Peer cell coverage: authored assessed 72, unassessed 0. Authored assessments are not verified outcomes.

Capability	Nishi	Bing-Videos	Yandex-Images	Google-Lens	Adult-DBs
Server-side clean delivery of a result page Source present; authored stronger claim: `cs_clean_page` in `runtime/_hdl_build/nx_clean_serve.nx`, verified at emit. LANDED (/compare/cleanserve). The result plays in a native player with the hostile page neutralized server-side; the engines iframe a provider or link out Adoption: `LIB-WIRED importers=3 nonval=2` — fully adopted (top of its ladder). Pros Source symbol found; stronger capability claim is authored; ahead of Bing-Videos, Yandex-Images, Google-Lens, Adult-DBs; fully adopted on the estate ladderCons none on the measured axes (rival marks are documented presence, not depth)	◉	○	○	○	○
Media resolver: declared metadata, page-id ranking, no ad or preview clips Source present; authored stronger claim: `csd_media_rank` in `runtime/_hdl_build/nx_clean_serve_daemon.nx`, verified at emit. LANDED 2026-09-02 (cleanserve CS9). The delivery half of this lane already refuses a related-video grid and strips ad videos Adoption: `LIVE-DAEMON` — fully adopted (top of its ladder). Pros Source symbol found; stronger capability claim is authored; ahead of Bing-Videos, Yandex-Images, Google-Lens, Adult-DBs; fully adopted on the estate ladderCons none on the measured axes (rival marks are documented presence, not depth)	◉	○	○	○	○
Reverse image - copy tier Measured: `s_verb_query` exists in `runtime/nx_imgsearch.nx`, verified at emit. LANDED. 992 permil over 17 transforms on /compare/mediaingest; Yandex Images and Google Lens are the mature bar rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). not adopted: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked) Pros Implementing source symbol found; runtime acceptance separate; ahead of Bing-Videos, Adult-DBsCons behind Yandex-Images (leads), Google-Lens (leads); not adopted yet: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked)	●	○	◉	◉	○
Identity tier of the semantic ruler Measured: `sr_hit_is_identity` exists in `runtime/nx_semruler_lib.nx`, verified at emit. LANDED AS A RULER, UNMEASURED AS A CAPABILITY: the decision core names the identity hit class and publishes permil -1 because no encoder feeds it. Listed so the P4 encoder rung carries its momentum rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them Adoption: `LIB-WIRED importers=2 nonval=1` — fully adopted (top of its ladder). Pros Implementing source symbol found; runtime acceptance separate; ahead of Bing-Videos, Google-Lens, Adult-DBs; fully adopted on the estate ladderCons behind Yandex-Images (leads)	●	○	◉	◐	○
Entity dossier from live evidence Measured: `ed_pick` exists in `runtime/nx_entity_dossier.nx`, verified at emit. LANDED and UNWIRED to any performer record; the databases are per-person by construction rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). not adopted: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked) Pros Implementing source symbol found; runtime acceptance separate; ahead of Yandex-ImagesCons behind Adult-DBs (leads); not adopted yet: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked)	●	●	○	●	◉
Seed discovery: one entity page to the others Measured: `es_has_tok` exists in `runtime/nx_entity_seeds.nx`, verified at emit. LANDED and UNWIRED; the engines' index is the bar rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). not adopted: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked) Pros Implementing source symbol found; runtime acceptance separate; ahead of Yandex-ImagesCons behind Bing-Videos (leads); not adopted yet: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked)	●	◉	○	●	●
Multi-image entity media gather Measured: `em_doss` exists in `runtime/nx_entity_media.nx`, verified at emit. LANDED and UNWIRED (javdatabase and wikimedia adapters); the reference-set rung composes it rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). not adopted: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked) Pros Implementing source symbol found; runtime acceptance separateCons behind Adult-DBs (leads); not adopted yet: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked)	●	●	●	●	◉
Entity-depth page ingest Measured: `pi_one` exists in `runtime/_hdl_build/nx_page_ingest.nx`, verified at emit. LANDED and UNWIRED; the corpus-growth rung drives it from the record rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). not adopted: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked) Pros Implementing source symbol found; runtime acceptance separate; ahead of Yandex-Images, Google-LensCons behind Bing-Videos (leads); not adopted yet: callable, authorised, completed MCP invocation not established by supplied journal window; lifetime history unknown (direct execution may be recorded elsewhere; this scan does not establish lifetime use); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked)	●	◉	○	○	●
Performer entity record and alias table Open — watching `runtime/nx_performer_lib.nx` : `pe_load`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P0. Bing resolves 14 of its 18 julia kyoka records through the alias Julia Boin and JAV codes; the databases hold them as fields; Nishi has no record rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching pe_load Pros none measured yetCons behind Bing-Videos, Google-Lens, Adult-DBs (leads); open contract; expected source declaration not established	○	◐	○	●	◉
Seed harvest through the existing organs Open — watching `runtime/nx_performer_harvest.nx` : `pe_seed_harvest`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P1. The organs exist; the record drives them under the mediaingest pacer rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching pe_seed_harvest Pros none measured yetCons behind Bing-Videos (leads), Adult-DBs; open contract; expected source declaration not established	○	◉	○	○	●
Entity-aware text ranking Open — watching `runtime/nx_performer_rank.nx` : `pe_query_expand`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P2. Baseline 0 relevant of 30 for both names against Bing's 54 and 18 rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching pe_query_expand Pros none measured yetCons behind Bing-Videos (leads), Adult-DBs; open contract; expected source declaration not established	○	◉	○	○	●
Documented reference set with provenance Open — watching `runtime/nx_performer_refset.nx` : `pe_refset`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P3. The databases hold it; Nishi has none rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching pe_refset Pros none measured yetCons behind Adult-DBs (leads); open contract; expected source declaration not established	○	○	○	○	◉
Sovereign face and image encoder Open — watching `runtime/nx_face_embed.nx` : `fe_embed`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P4, the encoder slot media R10 and mediaingest R17 share; the exceed is sovereignty, every rival's encoder is a hosted service rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching fe_embed Pros none measured yetCons behind Yandex-Images (leads), Google-Lens (leads); open contract; expected source declaration not established	○	○	◉	◉	○
Identity verdict with a referee panel Open — watching `runtime/nx_performer_identity.nx` : `pe_identity_verdict`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P5. Three outcomes, a reference floor, a calibrated threshold, an independent panel rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching pe_identity_verdict Pros none measured yetCons behind Yandex-Images (leads), Google-Lens; open contract; expected source declaration not established	○	○	◉	◐	○
Video and gif identity by sampled frames Open — watching `runtime/nx_performer_identity.nx` : `pe_video_identity`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P6. Nobody answers is-it-her for a video today; the exceed is real if the verdict holds dossierwatching pe_video_identity Pros none measured yetCons open contract; expected source declaration not established	○	○	○	○	○
Results filtered to the performer Open — watching `runtime/nx_performer_rank.nx` : `pe_filter`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P7. Bing mixes the artist and the footballer into the related list; the databases never do rival marks uncited — the Yes, Best or Part codes on this row are an observation read with no reference mark behind them dossierwatching pe_filter Pros none measured yetCons behind Bing-Videos, Adult-DBs (leads); open contract; expected source declaration not established	○	◐	○	○	◉
Referee census beats Open — watching `runtime/nx_performer_referee.nx` : `pe_referee_census`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P8. No engine measures itself against the others on a time spine dossierwatching pe_referee_census Pros none measured yetCons open contract; expected source declaration not established	○	○	○	○	○
Clean-serve as the delivery sandbox Open — watching `runtime/nx_performer_rank.nx` : `pe_clean_delivery`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. OPEN -- rung P9. Every result served through /clean with the relay; playback measured by the real-browser referee dossierwatching pe_clean_delivery Pros none measured yetCons open contract; expected source declaration not established	○	○	○	○	○

Rival-claim provenance. Every Yes, Best or Part mark in a rival column is a claim about someone else's product. A cited row carries a reference mark that resolves to a pinned mirror (the refs gate measures that); an uncited row is an observation read and is badged in its evidence until a reference lands. The badge is the receipt, never the proof: a mark proves a mirror exists, not that the mirror supports the code. The rows are counted below.

Rival claims, counted
Class	Count	Note
rows with rival marks	13	rows carrying a Yes, Best or Part in a rival column
cited	0	a reference mark resolves to a pinned mirror
uncited	13	an observation read, badged until a reference lands
rows, summed	13	RECONCILED: the parts sum to the total 13

Delivery and evidence

Inspect risks, technical debt, rendered observations, experiments and references. Read scope and limitations alongside every result.

Debt register

Id	Sev	What it is	Unblock
D1	8	No sovereign face or image encoder in NishiLang: the msm_search slot is published at permil -1 and every identity rung above depends on it	A ported encoder on the no-float seat with a KAT against a banked reference and a labeled pair set (rung P4); the no-float 7B OOM at 12 GB (debt 1788362867) bounds the seat
D2	6	The nx_cc module and closure limit: importing one more lib into a daemon tipped its closure past the limit (mediaingest R0 note), which is why the clean-serve daemon on the NAS cannot compose the resolver lib and the relay without the compiler lane	The compiler lane raises or partitions the closure limit; until then compose through small libs and measure the closure before importing
D3	7	No streaming HTTPS primitive: nx_https_fetch_follow buffers to a cap, so the CS7 relay (required for 4 of 5 KVS videos) and any video-scale identity sampling need either a windowed Range loop or a record-streaming variant of ff_core	A streaming client with Range and extra-header support in the fetch lib
D4	5	Own-verdict gates are refused promotion under D001, so executed evidence for clean-serve (52 of 52 locally) cannot be counted on the board	Migrate nx_clean_serve_gate onto gv_check (nx_gate_dry_apply then nx_gate_migrate verify)
D5	7	The Nishi text index does not know the entity: 0 relevant of 30 for both names against Bing's 54 and 18	Rungs P0 and P2; the index is fine, the ranking has no entity signal
D6	4	Google Videos and Lens answer a JavaScript shell to a non-browser client, so their referee rows are UNMEASURED from the NAS	The laptop Waterfox referee beat (nishi-ops ac_marionette.py) feeds the Google column
D7	4	Windows curl (schannel) fails the TLS handshake on some CDN nodes (2.prostoporno.place) where a browser may not, so a laptop probe can mis-classify a stream	Probe streams with the sovereign client (nx_https_get) on the NAS, never with schannel

Risk register

Risk	Likelihood x impact	Mitigation
Identity verdicts on adult performers are a privacy-sensitive capability	likely x severe	Reference sets come only from documented public sources with provenance rows, verdicts publish with the referee grade, and every record carries the operator's acceptance row before a performer is searchable
A wrong HER verdict attaches a person to content that is not theirs	possible x severe	Three outcomes with UNMEASURED as the default, a reference-set floor, a calibrated threshold on a labeled set and an independent referee panel before any filtering
The engines change their result shape and the referee beats go stale	likely x moderate	The beats parse the record fields measured today (vrhm murl, purl, mmsturl) and REFUSE with a named field when the shape drifts

Release history and work log

The worked plan, on the board. Every leg appends what it measured, landed, retracted, learned and left queued, per rung, so a crash leaves the next seat a ledger here rather than a transcript to mine. Kinds: measure, land, retract, lesson, queue. Newest last.

When	Rung	Kind	Entry
2026-09-02	P0	measure	BASELINE 2026-09-02 (this session): Nishi web search returned 0 relevant of 30 for julia kyoka (Penguin, Wikipedia Julia Scott, a German TV page, Lewd Swedes Julia Crown) and 0 of 30 for chloe lamb (Lemmy, an obituary, NASA, a handbag blog). Bing Videos: chloe lamb 54 records, 51 titles carrying both names, 48 on adult hosts (camwhores 13, tnaflix 12, peekvids 7, txxx 4, nuvid 4), 20 with a Bing-re-hosted preview clip; julia kyoka 18 records, 4 titles with both names, the rest by the alias Julia Boin and JAV codes REBD-425 and PPPD-335. Bing Images 34 and 100 tiles with Visual Search. Yandex Video about 17 each. Yandex Images CBIR present. Google Videos: a JavaScript shell to the probe. DDG 10 results. Estate incumbents: nx_entity_dossier, nx_entity_seeds, nx_entity_media (javdatabase adapter), nx_page_ingest exist and are unwired to any performer record; nx_imgsearch copy tier 992 permil; nx_semruler_lib sr_hit_is_identity declared, permil -1; nx_face_count cardinality only; no encoder.
2026-09-02	P0	queue	OPERATOR DIRECTION 2026-09-02: integrate clean-serve with the media capabilities so a known performer is searched by text and media, the corpus grows, the estate knows if it is her in video, image or gif like Bing does, results are only hers, and the engines are measured on /compare. Test entity: Julia Kyoka first (documented sources, javdatabase adapter, catalogue codes), Chloe Lamb second (the live clean-serve corpus). First-byte blockers named as debt rows D1 to D7 above.
2026-09-02	P0	measure	INCUMBENT MEASURED: nx_entity_dossier chloe lamb returned found=false and picked en.wikipedia.org Celtic_stone_idols as the landing page (card fields 0, authority 0, coverage permil 0 over 19,811 indexed matches) -- the dossier organ is real, runs, and has no performer record to anchor on; the julia kyoka run is in flight. The referee mirrors for the bar rows below were fetched through nx_research_fetch and pinned by hash.
2026-09-02	P0	measure	nx_entity_dossier julia kyoka: found=true on www.javdatabase.com (pick kind 2, card fields 8: JP name JULIA, DOB 1987-05-25, Tokyo, measurements, cup, height, hair; authority_total 1051 titles; codes sampled STOL-139, NHDTA-633) -- the estate CAN anchor a database-documented performer today, which is why Julia Kyoka is the first test entity and the P0 record starts from this card plus the alias Julia Boin the engines resolve through. Two defects measured in the same run: coverage.permil reads 18716 (indexed_matched 19671 over authority 1051), a ratio above 1000 that cannot be a fraction, because indexed_matched counts raw name hits across the index rather than entity-linked pages; and for chloe lamb the organ picked a Wikipedia page about Celtic stone idols with found=false, so its landing choice has no negative-alias guard. Both feed P0 and P2.

On these two registers. Rows are declared in the domain's plan file and carry the debt id, which is the join key back to the sovereign debt plane — that plane, not this page, is the authority on state. Reconciling them automatically (the regen reading the plane and refreshing these rows) is a named, owed rung; until it lands, treat an id here as a pointer to look up, not a status to trust.

SOTA field — 7 peers, 22 axes, measured bindings

Nishi (entity dossier + seeds + reverse-image copy tier + clean-serve delivery) vs Bing Videos, Bing Visual Search, Google Videos, Google Lens, Yandex Video, Yandex Images (CBIR) and the adult databases (IAFD, javdatabase, Boobpedia)

How the field is scored. A state-of-the-art comparison across the FULL competitor field: quantitative axes carry measured or published numbers (Nishi’s column is measured on this estate, competitor columns are researcher-sourced and cited in the note); grade axes carry Best / Yes / Part / No for every peer. A Nishi grade may be bound to an organ and a symbol and is then re-measured on every publish by the one symbol ruler the capability matrix above uses; an over-claim refuses to publish.

field, strip order:Bing-VideosBing-Visual-SearchGoogle-VideosGoogle-LensYandex-VideoYandex-ImagesAdult-DBs·BestYesPartNo

Text search

Name query returns the performer's videos

MEASURED: Nishi index 0 relevant of 30 for both names; Bing 54 records for chloe lamb (51 titles carry both names, 48 adult hosts) and 18 for julia kyoka; Yandex Video about 17 items; Google Videos returned a JS shell to the probe (documented capability, UNMEASURED here)

Alias and catalogue-code knowledge

MEASURED on julia kyoka: only 4 of Bing's 18 titles carry both names, the rest are reached through the alias Julia Boin and JAV codes (REBD-425, PPPD-335); javdatabase carries aliases and codes as first-class fields; Nishi has no alias table

Entity disambiguation - the performer, not the artist or the footballer

MEASURED on Bing's related list for chloe lamb: the artist Chloe Lamb (Vimeo), a footballer and an ASMR channel sit beside the performer; Google's knowledge panel disambiguates; the databases are per-person by construction; Nishi has no entity record

Entity landing page composed from live evidence

nx_entity_dossier exists (card, portfolio, coverage) and is UNWIRED to any performer record; Google's panel is the bar

PART

Media search

Reverse image - the same image and its transforms

nx_imgsearch copy tier measured 992 permil over 17 transforms (/compare/mediaingest); Bing Visual Search, Google Lens and Yandex Images are the mature bar

YES

Same-person retrieval from a face

honest gap: nx_semruler_lib defines the identity tier (sr_hit_is_identity) and publishes it UNMEASURED at permille -1 because no image encoder exists; Yandex Images CBIR is the documented leader for face lookup; Bing and Lens find lookalikes without an identity promise

Is it her in this VIDEO

nobody answers this for a video today; the Nishi path is keyframes through the first-IDR decoder into the same identity verdict (rung P6)

Is it her in this GIF or short clip

same gap as video; a gif is sampled frames

First-party preview clips and posters

MEASURED: Bing re-hosts a 1.1 MB video/mp4 preview per result on th.bing.com (CORS star, 30-day cache); Nishi's trickplay and sprite rungs are open (media R6, mediaingest R15)

Delivery

Plays the video on the result page

MEASURED: clean-serve emits the native player for every KVS page probed; in-browser cross-origin playback 1 of 5 until the CS7 relay; Bing plays only embeddable providers (iframe), Google links out, Yandex embeds many players

YES

No third-party JavaScript in the viewer

MEASURED 2026-09-02 on /compare/cleanserve (SANITIZE verdict, CS9 gate): clean-serve neutralizes every script server-side; every rival runs the provider's player or its own JavaScript

BEST

Safe ads preserved so the site earns

MEASURED 2026-09-02 on /compare/cleanserve (SANITIZE verdict, CS9 gate: safe ads kept while 728x90 autoplay ad videos are stripped and counted); no rival does this

BEST

Ad videos and preview clips never served as the content

MEASURED: 728x90 autoplay ad videos stripped and counted, preview clips refused, a related-video grid refused (cleanserve CS9)

YES

Corpus growth

Per-performer URL corpus that grows from use

knowledge/cleanserve_sites.conf holds 14 hand-seeded rows with measured classes; nx_entity_seeds and nx_page_ingest exist for growth and are unwired; the engines' index is the bar

PART

Documented reference set with provenance per performer

the databases carry the reference images and metadata with a source; Nishi has none; javdatabase adapter exists in nx_entity_media

Identity fan-out across sites by username

mediaingest R7 if_probe is the shared contract (data rows, paced); unbuilt

Sovereignty

Zero third-party runtime dependency

own language (nx_cc self-host, nx_cc_equiv_gate promoted), own TLS (nx_tls12_prf_gate promoted), own fetch and the clean-serve filter (CS9 gate, MEASURED 2026-09-02) and decoder; the encoder is planned and is not claimed here; every rival is a hosted service

BEST

Own index rather than a leased one

the Nishi index (crawl + Common Crawl) exists and does not know the entity; the engines' indexes are the bar

YES

Scale - quantitative

Video records returned for chloe lamb

measured 2026-09-02 with SafeSearch off

Bing-Videos 54Bing-Visual-Search n/aGoogle-Videos UNMEASURED (JS shell)Google-Lens n/aYandex-Video about 17Yandex-Images n/aAdult-DBs n/a

0 relevant of 30

Nishi, measured

Video records returned for julia kyoka

measured 2026-09-02; Bing resolves 14 of its 18 through alias and catalogue code

Bing-Videos 18Bing-Visual-Search n/aGoogle-Videos UNMEASURED (JS shell)Google-Lens n/aYandex-Video about 17Yandex-Images n/aAdult-DBs n/a

0 relevant of 30

Nishi, measured

Reverse-image copy tier

nx_imgsearch, measured on /compare/mediaingest

Bing-Videos n/aBing-Visual-Search documentedGoogle-Videos n/aGoogle-Lens documentedYandex-Video n/aYandex-Images documentedAdult-DBs n/a

992 permil over 17 transforms

Nishi, measured

Face identity tier

the sr_hit_is_identity slot is declared and empty: no encoder

Bing-Videos n/aBing-Visual-Search n/aGoogle-Videos n/aGoogle-Lens n/aYandex-Video n/aYandex-Images documentedAdult-DBs n/a

permil -1 UNMEASURED

Nishi, measured

7 competitors22 axes4 quantitativeNishi best 3 · absent 9

Honest verdict, measured 2026-09-02. Nishi is at zero on every identity axis: the text index returned 0 relevant hits of 30 for both names (books, obituaries and a NASA release for chloe lamb; Penguin, Wikipedia and a German TV page for julia kyoka), there is no face embedding, no is-it-her verdict, no per-performer reference set and no alias table. The sovereign wins are real but narrow: a copy-tier reverse image engine (nx_imgsearch, 992 permil over 17 transforms), a zero-third-party clean delivery that neutralizes the page and preserves safe ads, and the entity-dossier and seed-discovery organs that exist and are unwired to any performer record. The referees, measured live: Bing Videos returned 54 records for chloe lamb with 51 titles carrying both names and 48 on adult hosts, 20 with a Bing-re-hosted preview clip, and 18 for julia kyoka with only 4 titles carrying both names (the rest resolved through the alias Julia Boin and JAV codes REBD-425, PPPD-335); Bing Images 34 and 100 tiles with Visual Search; Yandex Video about 17 items each; Yandex Images with its CBIR face search; Google Videos serves a JavaScript shell to a non-browser client, so it is UNMEASURED here until the Waterfox referee beat runs. The backlog in its own order: an entity record with aliases and catalogue codes as data, seed harvest through the existing organs, entity-aware text ranking, a documented reference set, a sovereign face and image encoder (the same slot media R10 and mediaingest R17 need), an identity verdict with a referee panel, video and gif sampling, result filtering, referee census beats, and clean-serve as the delivery sandbox.

Honest verdict. The coverage above records source-symbol presence. Stronger comparisons are authored claims, not measured runtime superiority or accepted outcomes. A watch contract names an expected source symbol; finding it changes source discovery only. Qualification, deployment and runtime acceptance require their own evidence.

Person · product · place — not yet measured for this domain

Every compare carries this layer. Declare knowledge/compare/performersearch.ppp (rows surface|nishi or c1..c4|label|url|connect naming OUR live surface and each rival's front door), run nx_ppp_probe domain performersearch, and this section fills itself on the next beat: the same ruler on both sides — privacy and CX (third-party hosts, tracker classes, cookies, security headers), design and longevity (design hygiene, computed WCAG contrast, render-blocking resources, unsized media, script weight, theme and motion queries), findability (landmarks, skip link, on-site search, breadcrumb, headings, internal links).

The field — discovered, not chosen

Rows written by nx_field_discover from performersearch.seeds: the industry's own lists (Wikipedia wikitext, GitHub topics, awesome lists) read mechanically, every candidate counted across seeds. The matrix columns above are a SEAT'S pick; this band is the population they were picked from, and the stats line measures one against the other. A rival here is a lead, never a verdict — it earns a column when its capabilities are read and pinned.

Discovery receipt. One row per counter the run wrote, in the receipt's own order; the rival rows follow in their own table.

Field discovery, counted
Class	Count	Note
seeds	3
scanned	3
fetched	3
reused	0
failed	0
named	3
candidates	3
mentions	3
capped	0

Rank	Rival	Seeds	Mentions	First seed	Kind	Link
1	Bing-Videos	1	1	col0-bing	named	https://explore.microsoft.com/en-us/bing/visual-search
2	Yandex-Images	1	1	col1-yandex	named	https://yandex.com/images/
3	Google-Lens	1	1	col2-google	named	https://lens.google/

Exclusive and shared capabilities

One capability is graded on one board; every other board references it. The fleet duplicate census (nx_rivalcensus dupes, on the fleetdupes clock row) reads every board's matrix and names each capability graded as a row on two or more boards, with a home derived from evidence: rung when exactly one board's plan contracts the symbol (that board is starred), gates when no plan contracts it and exactly one board's executed gate map runs the organ's gate (that board is plussed), ambiguous when several boards carry the deciding evidence (a decision for their seats), and none when no evidence decides it, with the reason beside it: unmapped (no board of the duplicate declares a gates map; declare the maps) or unproven (a map exists and runs none of the organ's gates; gate the organ on its proving board, or declare it a shared library by policy). A home elsewhere means this board should carry an @xrow reference to the home board's row rather than a graded row of its own; a home here means the other boards should. The table is this board's slice of one estate-wide census: the shape is shared, the rows are the board's.

This board grades 18 capabilities: 16 exclusive to it and 2 shared with other boards. Shared rows on this page: 2 = home here 0 + home elsewhere 2 + ambiguous 0 + undecided 0 + refused 0 (partition 2); undecided = unmapped 0 + unproven 0 + unevidenced 0; homonyms 0; since the previous census: new 0, resolved 0. Census FRESH age_s=6152.

Capability	Organ	Boards grading it	Home	Rule	Evidence	Standing on this board
cs_clean_page	`nx_clean_serve.nx`	performersearch, cleanserve+	cleanserve	gates	gates	home elsewhere adopt: `@xrow cleanserve\|cs_clean_page\|home by gates (fleet duplicate census)`
csd_media_rank	`nx_clean_serve_daemon.nx`	performersearch, cleanserve*+	cleanserve	rung	plan	home elsewhere adopt: `@xrow cleanserve\|csd_media_rank\|home by rung (fleet duplicate census)`

Gaps from the record — what the estate does that no board carries

The record census (nx_goalmap record) reads the invoked-tool population and every plan queue row and files each organ or directive that NO matrix, plan or gates row names. A row here is a callout the boards missed: adjudicate it onto a board or declare it infrastructure. Census state STALE (age 1395257 s), sources read 4 of 7 declared — a BLIND census is a FLOOR: unread sources can only add rows.

kind	name	board	source	evidence

rows shown 0|this board's directives 0|estate-wide un-boarded organs 452 (listed in full on /compare/ecosystem)|census rows 639|malformed 0 (counted, never rendered)

References

Beyond a link list. Every reference below resolves twice — the publisher's copy and, where banked, the estate's own non-rottable library mirror with a content pin — and carries its evidence class plus the exact claim on this page it grounds. Keyed marks like [key] in the matrix notes jump here. A dash means honestly absent, never assumed.

[bing-vs] Microsoft Bing, Visual Search (reverse image by upload and camera), product page as served 2026-09-02 publisher · read in our library knowledge/fetched/cmp_performersearch_bing_visualsearch.html · pin hf43f6764d138812cab03a0432406e42c7893909121b8c4547a3ebf3cd5b9908d · accessed 2026-09-05 · vendor-docGrounds: Bing Videos and Visual Search
[yandex-images] Yandex Images, the image search surface carrying CBIR (content-based image retrieval) including same-face lookup, as served 2026-09-02 publisher · read in our library knowledge/fetched/cmp_performersearch_yandex_images.html · pin h6568b04eb956ae75fa1965229de37f4bc508114c6ffda7d04bade238286492c2 · accessed 2026-09-02 · vendor-docGrounds: Yandex Images CBIR
[google-lens] Google Lens product page (reverse image, lookalike retrieval), as served 2026-09-02 publisher · read in our library knowledge/fetched/cmp_performersearch_google_lens.html · pin h710c14d90ea2849fdbc61bd134aff6c1f0a82aa79a2a258b0a9c82d660208af0 · accessed 2026-09-02 · vendor-docGrounds: Google Lens and Google Videos

generated by nx_swcompare_matrix (sovereign NishiLang organ) from knowledge/compare/performersearch.matrix · source checks show implementation presence; runtime and user-outcome evidence are reported separately · JavaScript supports page controls