Next: 100.

• The delimiter-only lead and held empty interval.
• Shared travel, glyph entry and connector optics.
• The completed phrase and reading hold.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• The comparison reference is re-encoded. An independent original-master check found differences in 186/186 RGB frames and 186/186 native YUV frames. Pixel differences do not establish a perceptual defect or explain a prior preference. Native-source re-review is pending; supplied-file static preferences remain provisional.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• The range endpoints and growing highlight.
• Pointer, text and camera displacement together.
• The exposed workspace and retained selection.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• The comparison reference is re-encoded. An independent original-master check found differences in 144/144 RGB frames and 144/144 native YUV frames. Pixel differences do not establish a perceptual defect or explain a prior preference. Native-source re-review is pending; supplied-file static preferences remain provisional.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• The selected file count and drag-ghost detachment.
• Pointer and window occlusion during the crossing.
• Explicit editorial cue relocation, with the source overlap and unfinished outgoing context still documented.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• The comparison reference is re-encoded. An independent original-master check found differences in 124/124 RGB frames and 124/124 native YUV frames. Pixel differences do not establish a perceptual defect or explain a prior preference. Native-source re-review is pending; supplied-file static preferences remain provisional.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• The selected oversized layer and visible handles.
• Pointer, resized object, contact shadow and placement together.
• Source frame 1614, local frame 108: native bottom-boundary crops show lower sticker content outside the image in source and adaptation. Context checks show it fits at local frames 75 and 129. This intermediate excursion is not evidence of a broken mask or an approved reason to change the source choreography.
• The still-selected resize-ready ending; no invented deselection or completed settle.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• This lossless reference preserves the selected original-master RGB/YUV frames and native timestamps in the independent decoder check. That does not establish browser-pixel equivalence, moving perception, full-video coverage or source-fidelity approval of the adaptation.
• Measured source/output frame-position differences reach 0.625 milliseconds; frame-duration differences reach 0.0208333 milliseconds. Frame stepping uses separate native-clock tables. Matching positions do not erase or approve differing duration fields.
• The excerpt coded duration differs from the next source-frame boundary by -0.3125 milliseconds. Both measurements remain in the timing JSON.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• Stable outer-window geometry while space is allocated to a result pane.
• The blank wait, control states and result reveal.
• Small control contrast, text edges and ending context.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• The comparison reference is re-encoded. An independent original-master check found differences in 110/110 RGB frames and 110/110 native YUV frames. Pixel differences do not establish a perceptual defect or explain a prior preference. Native-source re-review is pending; supplied-file static preferences remain provisional.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• The identity position and optical framing on either side of the coverage change.
• The cut boundary versus movement within each view.
• Overlay contrast, background continuity and ending context.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• The comparison reference is re-encoded. An independent original-master check found differences in 147/147 RGB frames and 147/147 native YUV frames. Pixel differences do not establish a perceptual defect or explain a prior preference. Native-source re-review is pending; supplied-file static preferences remain provisional.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• Whole-word and glyph-built text arrivals.
• Independent line settling, overlapping actions and the common hold.
• Ordered text release over a continuing original background.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• This lossless reference preserves the selected original-master RGB/YUV frames and native timestamps in the independent decoder check. That does not establish browser-pixel equivalence, moving perception, full-video coverage or source-fidelity approval of the adaptation.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• Source frames [1640, 1658), 27.333333–27.633333 s: the near-stationary baseline before the first clearly visible horizontal displacement. This is a bounded still-frame observation, not proof of absolute rest.
• Source frames [1658, 1796), 27.633333–29.933333 s: leftward table translation and clipping inside the fixed phone aperture, with the heading anchored.
• Estimated internal phase markers: a renewed larger leftward step around frame 1752 (29.200000 s), then a far-left extreme around frames 1772–1773 (29.533333–29.550000 s) and a small rightward recovery. Gesture, physics and separate-input count are not established.
• Source return bracket [1796, 1911), 29.933333–31.850000 s: a larger decelerating rightward return to earlier columns. Inner phase boundaries are estimates, not exact input-event timestamps.
• Manual endpoint estimate: the table remains roughly 17 source pixels short of its starting position. This is not an exact restoration, measured zero-velocity hold or clean loop.
• Source frames [1774, 1780), local frames 134–139: native context checks show later product columns without the attribute-label column. Units make some values inferable; this is a reading-context limitation shared with the source, not proof of a masking bug. A retained semantic guide would be an editorial adaptation, not an observed source feature.
• The next clearly visible vertical displacement starts at source frame 1911 (31.850000 s), outside this selected excerpt. Its absence here does not establish a completed whole-film action or rule out other small movements.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• This lossless reference preserves the selected original-master RGB/YUV frames and native timestamps in the independent decoder check. That does not establish browser-pixel equivalence, moving perception, full-video coverage or source-fidelity approval of the adaptation.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• Source cut pair 506→507, 16.867000→16.900000 s: the browser is replaced by a white title screen already showing five glyphs. The unseen first five glyph additions must not be invented. The preceding frame is context, outside this excerpt.
• Source [507, 536), 16.900000–17.867000 s: first-line glyph additions accompany leftward displacement of existing glyphs. The complete first line is visible by frame 532; repeated wording does not imply identical pixels.
• Source [536, 564), 17.867000–18.800000 s: the second line begins with three visible glyphs while the first line rises. Further glyph additions and leftward travel overlap; exact settling endpoints are not recovered authoring keyframes.
• Source [564, 572), 18.800000–19.067000 s: the full phrase is present, but residual second-line travel and edge changes continue after the last glyph.
• Source [572, 591), 19.067000–19.700000 s: a near-stationary reading interval contains three separate exact-RGB runs, not one identical-pixel hold. Small changes between the runs are not assigned a rendering cause.
• Source [591, 597), 19.700000–19.900000 s: both lines enlarge slightly before the outgoing cut. The exact subpixel onset and easing remain unknown; the cut at frame 597 is outside the selected excerpt.
• The fitted title surround and inspected native patches appear white. Uninspected native regions and faint measured residuals do not establish a separate blur trail, mask or wipe. Geist outlines and replacement copy remain editorial choices.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• The source coded tail is five ticks beyond the next-frame boundary at a 16,000 Hz timescale. Source positions, fixed packet-duration fields and output H.264 duration normalization remain separate measurements.
• The builder replay uses pinned local dependencies and a disclosed local-only runtime workaround. It is not a clean-machine installation check.
• This lossless reference preserves the selected original-master RGB/YUV frames and native timestamps in the independent decoder check. That does not establish browser-pixel equivalence, moving perception, full-video coverage or source-fidelity approval of the adaptation.
• Measured source/output frame-position differences reach 0 milliseconds; frame-duration differences reach 0.6875 milliseconds. Frame stepping uses separate native-clock tables. Matching positions do not erase or approve differing duration fields.
• The excerpt coded duration differs from the next source-frame boundary by 0.3125 milliseconds. Both measurements remain in the timing JSON.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

• Source 2100→2101, 70.070000→70.103000 s: the first added code row, explorer folders and corner badge become discernible together. Sub-frame onset and exact opacity are not measured.
• Subsequent rows first appear at source frames 2113, 2123, 2133, 2143, 2153, 2163, 2173, 2183, 2193, 2203 and 2213. The first gap is 0.400000 s; subsequent native-PTS gaps vary between 0.333000 and 0.334000 s. These are whole-row appearances, not observed character typing.
• Source 2108→2109, 70.337000→70.370000 s: a new footer caption becomes discernible after a caption-free interval. Its estimated settling near frame 2118 is not an exact transform endpoint.
• Source 2132→2133, 71.137000→71.171000 s: a blue emphasis background and left accent appear with the new row. Later rows extend the lower edge while the caption remains present.
• Source final-row bracket 2212→2213, 73.807000→73.840000 s: the outer closing brace appears. The core looks settled by frame 2217, an estimate; this excerpt ends at 2220, not at the later caption exit.
• No pointer, caret, scrolling, broad reframe or cut is apparent in the fitted field and four inspected native regions. The remaining native area is not certified free of micro-events. The caption exit and following caption occur outside this excerpt and remain separate source-context observations.

• Native geometry and corresponding frame counts match. Each movie retains its own PTS; that does not establish source fidelity or resolve a clock discrepancy.
• Full-video coverage and moving-source review are not complete. Native PNG availability is not evidence of human temporal inspection.
• The source passage may begin or end within a larger action. Do not infer an unseen anticipation, settle or outcome.
• Original text, drawing, font and content choices are adaptations. Source assets and full downloads are not included.
• Publication is August 6, 2026. Metadata was retrieved on September 5, 2026 at 17:15:49 UTC, 128.397967 seconds after the recorded 17:13:40.602033 UTC cutoff; this retrieval qualification remains unresolved for collection sign-off.
• The lossless reference preserves original 533-tick packet durations. Its coded tail is 11 ticks before the next-frame boundary; the outputs preserve positions but normalize duration fields differently. The lossy convenience preview is not used as the comparison reference.
• This lossless reference preserves the selected original-master RGB/YUV frames and native timestamps in the independent decoder check. That does not establish browser-pixel equivalence, moving perception, full-video coverage or source-fidelity approval of the adaptation.
• Measured source/output frame-position differences reach 0 milliseconds; frame-duration differences reach 0.6875 milliseconds. Frame stepping uses separate native-clock tables. Matching positions do not erase or approve differing duration fields.
• The excerpt coded duration differs from the next source-frame boundary by -0.6875 milliseconds. Both measurements remain in the timing JSON.

Selected code/SVG candidates with original content; public package normalization remains a separate gate. The source excerpt is silent.

Observed differences between indexed stills, not a moving-footage review. Context rows extend beyond the delivered source parts and link to the official video. Estimates and unresolved boundaries remain explicit.

• Single white frame after the incoming cut.
• Faint central card before the yellow image resolves.
• Delayed right-card appearance and a later narrow edge sliver.
• Lateral travel, clipping and the central image change.
• Continued contraction through the final indexed frame, before the next source cut.

• The four source files are consecutive delivery parts of one passage, not four source shots or four treatments. Playback is manual between parts; no seamless-playback claim is made.
• The source is split to retain faint pixel states without exceeding the existing 8 MiB movie limit. Each RGB and YUV frame is compared against the local source master; this is not visual inspection credit.
• Lossless decoded files do not guarantee identical browser display values. The builder measured a faint right-card minimum of RGB 253 in the native source and 254 in Chromium at source frame 2767; the color-conversion cause is unresolved.
• The next source cut interrupts the shrinking cards. Do not invent a completed settle, hidden anticipation, easing curve or separate blur/opacity measurement.
• The lamp and vessel drawings, copy, typography and materials are original code/SVG adaptation choices, not OpenAI source assets.
• The source wins both first static comparisons. The second image remains under-resolved in the candidates after third-card entry, especially local frames 82–84 (source frames 2815–2817). A third-version repair awaits a new independent comparison; the reviewed first version remains visible here. Moving perception, full-video coverage, fidelity and completion remain unverified.
• The first extracted-package replays retain exact clocks but differ from the selected outputs at 22 Fold Light frames and 8 Kiln Vessels frames. The coordinator independently confirms those differing frame indices. Reproducibility remains under repair; this package has no pixel-identity approval.
• The first archive also fails three existing normalizer checks concerning local references, unavailable provenance and a checksum filename. A third-version archive repair has exact replays, but this gallery retains the reviewed first version and its failed checks. Its download remains withheld.