OpenClinXR

Source of truth for status is the repo ledger (PROJECT_STATUS.md), not this page. Public copy here is a snapshot for humans—updated when the product story actually moves.

Nine cards landed against a frozen acceptance contract. Each carries an evidence report graded by its own verifier against retained artifacts, and each was re-measured by the owner against the tree rather than accepted from its report. Three of nine were rejected at least once; one was rejected three times. The rejections and the failures are published alongside the passes, because a report that only lists passes is not evidence.

The last row is the useful one. Identifying a foot-contact window at the rubric's 0.005 m per-frame allowance needs roughly 132 Hz; four skinned humanoids and a compiled room through software rendering gave 8.00 Hz. Rather than grade an aliased stream, the instrument refuses and reports the number. An earlier revision of that report graded it as failing from an assumed 60 Hz that was never measured; that verdict was wrong in the same way a passing one would have been, and is retracted in place in the published report.

Other findings worth naming: the shipped locomotion clip fails the frozen rubric, and no clip this pipeline has produced meets the foot-plant threshold. A learned-motion candidate was screened and held, not adopted — its text encoder depends transitively on a gated model licence this project does not hold, and the host has no CUDA device. Twenty-eight open items are declared across the nine reports.

Reports: SC-00 rubric · SC-01 · SC-01S · SC-02 · SC-03 · SC-04 rights · SC-05 approach · SC-06 replay · SC-10 research hold

notEvidenceFor: clinical validity, scoring validity, exam equivalence, and worn-headset readiness are not claimed and have no evidence here. Three of twelve acceptance rows — an uninterrupted workflow recording, a website demonstration, and independent final acceptance — are unstarted, so this package is not closed. No humanoid imagery from these cards is published: their public render rights are blocked pending a named upstream licence resolution.

These screenshots were captured from commit 05120405, where BothyBoard children W1–W19 (plus W14a/W14b/W11s) had landed. Faculty author the encounter as a Scenario, then review a World Compile Graph of baker families. Add node / Remove node mutate the graph; the lock table stays the lock write path. Chromium Playwright of the running admin app (ui-admin + local API), not a schematic. Since this capture, authored revision identity now invalidates stale approval, and Admin can preview the runtime delta before a reviewed revision is promoted into an immutable encounter bundle.

claimScope: local admin UI capture of an authored Scenario at capture commit 05120405 plus the compile DAG recaptured 2026-09-10 at bbf3f84a; current revision identity and runtime-delta behavior are supported by later committed application tests, not by these stills. notEvidenceFor: Quest readiness, clinical validity, live Blender bake, LLM draft quality, or that the parent BothyBoard program card is closed (it is still Idle). Node labels truncate in this crop; that is the canvas, not missing data.

Each production factory station is a Standard Schema V1 interface (~standard.validate plus jsonSchema.input). Admin cards are derived from that schema: one card per id, one control per property, Apply refuses invalid payloads. instrument is a gate and is not a card. The TRELLIS control adds a worldview bake model (subject/pack ecg-cart-imagine-box), distinct from a fixtureSlot bind.

claimScope: local admin React cards derived from factoryStationSchemas on origin/main. notEvidenceFor: live TRELLIS GPU bake, Quest readiness, clinical validity, or that Apply invokes a baker (it validates and emits the payload).

Three of the fourteen stations lay their patient down. For weeks those three rendered a crumpled knot on the mattress. Five separate fixes landed and none of them changed a pixel — each was aimed at a cause nobody had located. On 2026-08-20 the sixth worked, because an ablation was run first instead of a sixth guess: render the pose's two mechanisms separately and see which one is the damage.

The cause was seventeen joint rotations hand-tuned against Anny's 23-bone skeleton, being applied to an MPFB body with 138 joints. They are correct for the rail they were written for, so the fix was not to delete them — it was to scope them. Anny keeps its table; MPFB skips it. One switch, and the right-hand cell above became the centre cell.

The part worth keeping is the measurement that could not tell them apart. Before the fix, the knot and the person had identical envelope metrics — height 446 mm, minimum Y 0.570, to three decimal places. One was a patient and one was refuse, and every bounding-box assertion in the repository scored them the same. Any future contract that grades a lying figure by its bounding box is measuring nothing.

Not fixed, and visible in that capture: her arms rest raised in the air rather than at her sides. That is the bind pose showing through now that the wrong rotations are gone, and the answer to it is a retargeted recumbent motion clip — not a second table of hand-authored angles. That work is open. Nothing here is a claim about clinical validity of the pose; it is a claim that a figure on a bed reads as a human being.

A case definition describes a person in numbers: age, height, build, how tense they look. Until today the factory did not read them. Bodies were baked from hand-written Python presets — four of them, covering four actors out of a hundred and four. Every other figure a learner met was the same body with a different job title, and the numbers in the case were decoration.

The case now authors the phenotype and the generator resolves it. Measured in the shipped fixture data: 32 of 32 actor records carry numeric phenotype, where before the audit found thirty-eight of forty-two actors with none at all.

patient_robert_hayes_v1
  age            52
  height_cm     178
  bmi            26
  brow_tension  0.55
  anxious       0.65
  flush         0.15

What this is not: it is not a claim that every shipped body has been re-baked from these numbers. It is the data and the seam that reads it — the resolver now has something to resolve. Re-baking the cast against it is the next piece of work and it is open. Nothing here is a claim about clinical validity of any figure; it is a claim that the case definition finally reaches the body.

A case definition describes a person: age, build, hair colour, eye colour. The factory was reading almost none of it. Every shipped figure's iris came from a role default — patients brown, family green, nurses blue — assigned by matching words in a job title. The case's own eye_color field was carried all the way to the bake and then dropped, because the one call site passed an empty dictionary where the phenotype belonged.

The more useful half is what happens when a case asks for something the factory cannot build. One patient is authored with hazel eyes. There is no hazel material in the licensed set, and the old behaviour was to silently fall back to a role default — the case asked for one thing, the learner saw another, and nothing anywhere said so. The factory now refuses that value out loud, and publishes the nine colours it can actually produce, each with its licence, generated from the same list the renderer reads so the two cannot drift apart.

That refusal is deliberate and it is the point. A factory that quietly substitutes is a factory whose output nobody can trust to match the case; a factory that refuses forces the question upstream, where a human or an authoring tool can answer it against a real list of options. The hazel patient is still unbuilt, and that is the correct state until someone chooses.

Not claimed: that these faces look finished. The skin is procedurally baked and carries no painted facial detail, which is visible in these crops and tracked separately. This shows one field of the case definition reaching one property of the render, verified by comparing the shipped bytes and by looking at the pixels.

Shipped MPFB bodies carry MakeClothes library garments fitted onto the standard rig: teal scrubs on the nurse, a cyan exam gown on the adult patient, sage street clothes and boots on the walk-in, muted rose on the family partner. Contrast against skin is visible at a glance. These are isolated grade captures of the tracked GLBs under apps/ui-xr/public/generated-humanoids/, not the 2026-08-21 sage t-shirt pair that documented a palette bug.

Street still: isolated EEVEE finished_figure_grade.py --frame full with world Background strength 0.55 (grade-lighting.json), 2026-09-10. Other cells: model-vetting-glb-grade-capture front_lit, 2026-08-25/28. Garments are fitted via MakeClothes ClothesService.fit_clothes_to_human onto the MPFB basemesh. A 2026-08-21 incident (cream closed_casual tint 21 RGB units from skin) is closed.

Still visible and not claimed as production cloth: nurse and family still show a waistband gap (cargo cover shell). Street trousers are MakeClothes pants02 punkduck_male_classic_jeans (pack-page CC-BY), not a cover shell. Isolated EEVEE bake includes world Background strength 0.55. Poke-through at the chest on some shirts. Isolated harness, not a station. claimScope: role-distinct fitted clothing on shipped MPFB GLBs. notEvidenceFor: production cloth, Quest readiness, pregnancy morphology, clinical validity, or exam equivalence.

On 2026-08-11 a rig upgrade landed on the library bodies — MPFB's 64-bone mixamo_unity skeleton and its shipped CC0 weight map, replacing a hand-rolled bounding-box armature whose hands carried 0.00% of the skin weight with 20,216 vertices collapsed onto a single upper-arm bone. Three machine contracts passed. Then the capture was graded by eye, and the figure had no trousers.

The revert took one commit. What it exposed took the rest of the day: the trousers were never rebuildable. Scrub_Shirt.mhclo and cortu_cargo_pants.mhclo were on no disk in the repository — the staging directory is empty and gitignored, and the provider cache held three upper garments and no lower-body source at all. The 2,530 triangles in the tracked asset came from a bake whose input no longer existed. Meanwhile the pipeline's own “find-or-stop” guard, asked for a lower garment and finding none, logged a warning and baked the body anyway.

Both halves are fixed and in the repository: the guard now throws instead of warning, and the sources are acquired, tracked, and recorded in the licence ledger at acquisition time — makehuman-pants01 (CC0, Cortu Johnstone) and Scrub_Shirt (CC-BY, WojackOWL). A re-bake now finds its inputs with no network, or refuses loudly.

What this picture is not. It is not a claim that the figure looks good. The same capture carries three defects we have measured and not fixed: a sawtooth band of bare skin where the shirt hem meets the trousers; hands painted the garment colour rather than clothed — the shirt mesh spans Y 0.913–1.505 m on a 1.760 m body while the hands sit near 0.79 m, below its lowest vertex; and a shirt with no volume of its own, breast and navel anatomy reading straight through it.

A correction we made to ourselves. Grading the lit pass alone, we recorded the hands as “mittens with no separated fingers” and were about to file missing hand geometry. The structure pass shows fingers and thumbs present and well formed. Lit resolves silhouette, structure resolves topology, and the lit pass flatters. The wrong finding was caught because both passes are captured, not because anyone was careful.

Captures produced by model-vetting-glb-grade-capture, whose NodeIO-versus-scene-graph self-check agreed to 1.2 × 10⁻⁵ relative error. That agreement proves the renderer drew the file and nothing about whether the file is right — which is why a human graded the pixels, and why the missing trousers were noticed at all.

The two library bodies were bound to a hand-rolled bounding-box armature with Blender's automatic weights. Measured, that put 0.00% of the skin weight on hand.L/R and collapsed 20,216 vertices onto a single upper-arm bone — an arm that moved as one rigid piece from shoulder to fingertip. The fingers were in the mesh. Nothing could move them.

They now ride MPFB's 64-bone mixamo_unity rig with its shipped CC0 weight map — mixamorig:LeftHand alone carries 592 vertex entries, with full finger chains. Both files shipped with MPFB and neither was being used.

Why supine, and why that is the better evidence. A standing rest pose looks identical before and after a skeleton change — it would prove nothing. These are recumbent renders from the isolated posture harness, and the folded arms and separated fingers are shapes the previous rig could not produce at all.

What these are not. Isolated lab renders, not in-station frames: no room, no other actors, no clinical context. Framing is computed from the subject's bounding box by the harness, not authored — which is the point, because the authored per-mode cameras in the full scene are a separate and still-open defect. Two attempts at hand-tuning those camera positions today were measured and reverted, one of them producing a 7.4 KB blank frame.

Defects still visible and unfixed: the sawtooth seam where the top meets the trousers, and low-polygon faceting across the limbs. Neither is claimed as solved.

Last-resort factory when the object is not in the kit/parametric store and not acquirable CC0/CC-BY. One Grok Imagine upper-¾ on a black void, flood-keyed (Imagine writes JPEG RGB — no native alpha), then TRELLIS.2 Metal with Space-order remesh on compact extracts, then factory:trellis:optimize to the preferred ≤80k stop. Six subjects ran that path. The 49M medication cart is a recorded miss. Kit remains exam SSOT. These GLBs are harness champions — not promoted into ui-xr.

Orchestrator pixel grade, 2026-08-14: four Imagine plates are isolated black-void product shots; four 80k champions are the same silhouettes with remesh-soft surfaces. Pulse-ox and glucometer also ran the hatch to 80k; their grade crops are either too tight (hinge cavity) or flank-on, so they stay in the ledger and off this page. claimScope: escape-hatch factory stills for compact hard-surface props. notEvidenceFor: Quest readiness, clinical accuracy, device equivalence, kit replacement, UI-XR promote, Imagine-shader smoothness. CLI: pnpm factory:trellis:hatch (commit c758a276). Evidence: .openclinxr/evidence/trellis-escape-hatch/ (gitignored champions).

The next work is to turn the newly connected lifecycle into visible, measured product evidence. Dates are not SLAs, and nothing here promises production deployment, headset certification, exam equivalence, or clinical validation.

Public roadmap copy follows committed product evidence and the repository’s protected claim boundaries. The operational queue remains in PROJECT_STATUS.md and the project board.

The repo runs an OpenClaw-style operating model: role charters, path scopes, leases, drift guards, and slice records. That keeps long autonomous work from inventing features outside the blueprint-factory mission. It is how we build OpenClinXR—not a separate product you install.