The whole project.
One place.

Square and portrait players, chapter seeking, original F1 comparison and editable-file links. Scratch narration; listening and final voice remain open.

1080×1080, 24 fps, 720 frames; real Rands footage, preserved illustrative claims, original sound sketch and temporary supplied narration.

1080×1920, separately composed layout; complete CTA holds from 26 to 30 seconds. Review deliverable, not final campaign approval.

Technical decoding and sampled native export inspection; voice and continuous listening/playback limitations.

Source preservation, exact export hashes, dimensions, frame counts, native observations, pending voice and client-reply state.

Seven timed beats from identity through growth, complaint, work/evidence, review and CTA; provenance, typography and two formats.

Local FFmpeg/Pillow project; source windows, motion, font fallback, rebuild commands and exact rendered snapshots.

Current source selections, placements and provenance. Source footage runs at natural speed.

Actual local composition and motion source used for the review cuts.

Preserved timeline input used by the delivered renders.

Preserved renderer input used by the delivered renders.

Source identity and selections, fonts, brand assets and public product crop.

Unrecorded 401-character script. This is not the voice in the current MP4s.

Pricing-only Higgsfield text2speech_v2 / elevenlabs receipt; no job submitted, no voice selected, spend confirmation pending.

Original DSP music/cues, temporary corrected F1 voice, ducking, measurements and unperformed direct listening.

Stereo 48 kHz review mix; no new speed or pitch change applied to supplied narration.

Corrected supplied F1 voice retained as scratch; historical compression is not undone.

Deterministic DSP instrumental sketch; requires listening acceptance.

Four original DSP accents for the authored beat changes.

Separate ducked instrumental/cue stem for remixing.

Deterministic oscillator/noise source, envelopes and mixing logic.

Both muxes measure −16.10 LUFS / −1.52 dBTP; numerical verification is separate from listening.

16 timed frames decoded from the actual final square MP4.

16 timed frames decoded from the actual final vertical MP4.

Caught newly introduced ceiling-damage and pre-work assertions, notes-image gap and source-identity wording.

Records the corrected illustrative text and source-work image before export.

Grouped source for timed proof images in both formats.

Why the selected vendor illustration fits the evidence story; it is not a Rands account capture.

Source image for the right-phone crop used in the ending, labeled Product example.

All five original sources, file hashes, measured media, outer/nested ZIP contents and missing promised production layers. Early perception status is historical.

32,552-byte recovered attachment; explicitly a reconstructed August history, not a verbatim conversation log.

28.468-second linked portrait cut; corrected combined-experience narration and real customer opening.

96.502-second 1280×720 landscape recording with two interview representatives and real HVAC/phone work.

29.985-second, 1320×1300 screen recording; quality/storytelling reference. Encoded streams match the earlier PromptLab source.

Original package: 39 real outer files plus 37 nested files, including statics, strategy, six tools and 13 brand PNGs.

28.958-second alternate with warehouse/office opening and nineteen-years narration; distinct from linked corrected F1.

Second 28.958-second ZIP encode; same sampled edit as ZIP portrait, not the current review cut.

Exact archive membership; distinguishes real files from macOS sidecars.

Grouped inventory of strategy, brand, research and scripts; promised separate video layers are absent.

Integrity evidence for all 76 real outer-plus-inner members.

Campaign evolution, source roles, Jobber/F1 comparison and educational preservation. Its assessment-only completion wording predates the current production cuts.

Nine strategy documents, six historical scripts, supplied transcripts, customer facts, invented scene origins and real code limitations.

Beginning-to-end sampled review, source people/work, counters, CTA holds, alternate cuts and corrections to local model output.

Direct authorized inspection of source frames: business history, damage proof, product record, review/brand overlap and look differences.

Two interview setups, roof/phone footage, source timing and visual differences between F1 versions.

Complete local ASR and signal measurements; no direct listening. Distinguishes corrected final from older nineteen-years versions.

Narrative meaning, timing, source support and the boundary between transcription and hearing.

PromptLab, Tarot, Rogowo, Stół, Memoirs and OSHA4US findings converted into practical source, motion, typography, sound and final-file checks. Preflight-era decisions remain historical.

Matching title/duration metadata and one thumbnail; no complete public-video identity or direct customer-review authorship established.

Original claims, disputed quotes, generated scenes, criticism and revisions retained with source provenance. No model training occurred.

Full seven-beat original, nineteen/combined-thirty-year variants, growth, complaint, readings, timestamp and review alongside ASR/OCR evidence.

Origin and support are separate labels; missing support is not proof of falsity. Original customer-review URL/authorship remain unverified.

Source → attempt → exact statement → evidence/provenance → critique → optional revision for each case.

Machine-readable row contract for the preserved educational cases.

18 valid schemas, 82 reference occurrences across 19 distinct files, 53 exact excerpts; no training executed.

Checks schema, evidence-file hashes and exact source excerpts without a model call.

Source preservation, native review evidence, contracts and final dataset validation before the subsequent production stage.

Actual extraction, historical prompt adaptation, parser cleanup/validation, applicable defects and skipped routes.

261 keyframes, three audio files, 27 historical structural recipe passes and 162 dialect strings.

Equal encoded video/audio payloads in different containers; enables evidence reuse without claiming file identity.

Five copied analysis/log baselines with hashes; past matching reached only 15/27 stills at its threshold.

Seven optional script beats, two fresh visual descriptors/recipes and four dialect strings through unchanged existing contracts; no model-driven reverse/generation.

Actual unchanged PromptLab script validation, canonical normalization/adapters and Reverse parser/cleanup/validation, with tool-source hashes.

Two source-based visual anchors validated by the existing contract; structural validity is not creative acceptance.

Four format adaptations; dialect names are not model selection or provider-job receipts.

Rands business-context image and Jobber evidence-visibility anchor with methods and hashes.

This exact reviewed body was sent 10 September 2026 at 18:48:36 UTC. No reply at the last recorded check at 21:22 UTC.

Source recovery and sent questions. Its production-on-hold and no-fresh-vision state were superseded by later authorized native review and production.

Historical commercial hold/exclusion decisions. Later educational steering preserves unsupported original examples with provenance.

Optional professionalism/consolidation branch: 49 narrator words + 14-word recording excerpt. It does not replace the preserved original or current production.

Dated snapshots and hashes from before educational-preservation steering.

Superseded production brief before the inspected public product-example crop was added.

Fresh local Whisper transcript of the complete audio track; estimates and uncertain spellings preserved, not a listening verdict.

Fresh local Whisper transcript of the complete audio track; estimates and uncertain spellings preserved, not a listening verdict.

Fresh local Whisper transcript of the complete audio track; estimates and uncertain spellings preserved, not a listening verdict.

Fresh local Whisper transcript of the complete audio track; estimates and uncertain spellings preserved, not a listening verdict.

Fresh local Whisper transcript of the complete audio track; estimates and uncertain spellings preserved, not a listening verdict.

Grouped native audio, 15-second review windows, source hashes and timebases for all five versions.

Loudness, peaks and quiet-interval measurements; does not identify music or judge vocal performance.

Timed beginning-to-end sampled boards for source/reference/attempts, including terminal gaps missed by initial keyframes.

48 source candidates for finding real interview, roof, phones and ending material; candidates may span cuts.

350% settled by 8 seconds; corrects a proposed late-counter diagnosis without substantiating the growth claim.

Complaint OCR and evolving dollar counter; $1,014,106 is an intermediate sampled value.

125 ms sampled text checks show full original ending around 27 seconds, leaving about 1.47 seconds before EOF.