ARTICLE 4 · FIND YOUR VOICE → GPT → LEONARDO → QA

The exact ChatGPT → Leonardo Seedance 2.5 workflow behind The Match

The useful lesson from The Match is not one magic prompt. It is the development path: human decisions first, ChatGPT turns those decisions into production instructions, Leonardo executes them, QA diagnoses the right layer, and controlled post-production finishes the approved result.

The jobs are different

Creator: decides meaning, audience, message, emotional goal, brand rules and what must remain human.

ChatGPT: structures those approved decisions, identifies missing decisions, creates the reference-image brief and translates the creative brief into a platform-specific prompt.

Leonardo: renders the image/video from the supplied reference, model settings and execution prompt.

CapCut Pro: will finish and package the approved visual for publishing without asking the generation model to control exact typography, series labels and final editorial timing.

Do not ask the production tool to solve a Voice problem that should have been decided before generation.

1 · Start with the Find Your Voice brief

Before writing a Leonardo prompt, answer these three questions:

What should this make people feel?
What should they understand or remember?
What makes this recognizably mine?

For The Match, the answers became:

FEEL: quiet recognition, warmth, creative connection and the moment something finally comes alive.

REMEMBER: creators can generate many things; the breakthrough is when something begins to feel like yours.

MINE: Bee Righteous + Match, black-and-gold visual language, restrained emotional storytelling, and the Jack Righteous line “Create What You Love. Love What You Create.”

2 · Give ChatGPT an instruction framework before asking for a prompt

The GPT should not be rewarded for filling every blank. It should protect the creator decisions already made.

CORE GPT INSTRUCTION

Work only from the creator decisions I provide. Preserve approved meaning, audience, message, tone, character rules and brand constraints. Do not invent missing identity decisions. If something important is missing, mark it [CREATOR DECISION REQUIRED] and ask one concrete question. Separate creative decisions from technical execution. When the creative layer is approved, translate it into a platform-specific production prompt without changing the underlying intent.

This instruction works beyond Bee Righteous. The point is to stop AI from quietly changing the project while making the prompt sound more polished.

3 · Build the reference before the video prompt

Our first character bible was useful for human design review but too busy as a production reference. It contained multiple poses, expressions, labels, diagrams and repeated versions of the characters.

The correction was a single clean production reference:

square composition
Bee Righteous + Match only
full-body
three-quarter front angle
same baseline so the size relationship is obvious
neutral expressions
plain warm neutral background
no props
no alternate poses
no environment
no readable text
no logos
only Match’s restrained chest glow as a character cue

Production principle: a character presentation board and a model reference image are not the same asset. The production reference should remove ambiguity.

4 · Leonardo setup used for refinement

SETTING	VALUE	WHY
Platform	Leonardo AI	Visual generation environment used for this run.
Video model	Seedance 2.5	Model used for the 15-second short.
Duration	15 seconds	Enough room for five clear emotional beats without padding.
Aspect ratio	9:16 vertical	Designed for short-form vertical delivery.
Refinement resolution	HD 720×1280	Used while refining the creative result before spending more on final delivery.
Audio	ON	Tests ambience, silence, chime and emotional bloom with the visual.
Private Mode	ON	Keeps the working generation private.
Legacy	OFF	Uses the current generation path rather than legacy behavior.

Leonardo also offered Standard 496×864 and Full HD 1080×1920. The refinement run stayed at 720×1280. The post-production master will use a 1080×1920 vertical canvas in CapCut.

5 · Translate the Voice brief into a platform prompt

The prompt was ordered by production priority rather than written as one giant descriptive paragraph: character lock → style/emotion → story priority → timed beats → camera → motion restraint → consistency restrictions → audio → final delivery priorities.

Longer is not automatically better. Once the creative decisions were settled, the production prompt became more directive.

6 · RUN 1 — FIRST DRY RUN

The first run proved that the idea was worth continuing.

What worked

the core emotional concept was readable
Bee Righteous remained recognizable
Match had charm
the discovery/reveal worked

What drifted

the final third became too visually busy
generic glowing music imagery weakened the specific world
the ending drifted toward a world-within-world transformation
Match’s spatial relationship became less clear
the fingertip moment was not deliberate enough

RUN 1 LESSON

Do not reject the whole project because one production layer drifts. Preserve what worked and identify the layer that failed.

SHORT 1 VIDEO — TO BE ADDED
The YouTube Short will show the first dry run and the specific drift we diagnosed.

7 · RUN 2 — CONTROL THE DRIFT

The biggest improvement did not come from simply adding more prompt language. We improved the input and simplified the execution.

What improved

clean canonical two-character production reference
stronger character consistency
clearer Bee / Match distinction
same studio retained
fewer effects and less spectacle
better recognition beat
improved emotional chemistry
cleaner ending

What still needed correction

the opening read too static and too sad rather than creatively dissatisfied
Match appeared too early / popped into view instead of being physically uncovered
the fingertip connection was still not unmistakable
the final hero frame was not deliberately held long enough
repeated bee drawings on the desk became too self-referential

SHORT 2 VIDEO — TO BE ADDED
The second YouTube Short will demonstrate how the cleaner reference and tighter instructions changed the result.

8 · RUN 3 — PRODUCTION RUN

The production direction keeps the same creative concept but makes causality and timing much harder for the model to misunderstand.

0–4 SEC · CREATIVE OVERLOAD

Bee is thoughtful and dissatisfied, not sad. The workspace shows real creative activity without turning into visual clutter.

4–7 SEC · TRUE HIDDEN DISCOVERY

Match remains completely hidden until Bee physically moves the final object and reveals Match. No pop-in.

7–10 SEC · HELD RECOGNITION

Mutual eye contact, tiny mirrored head tilt, flame brightens, Bee smiles. No room transformation yet.

10–12 SEC · UNMISTAKABLE CONTACT

Both fingertips stay visible in the same frame. Touch is held briefly. The gold pulse starts exactly at contact.

12–13.5 SEC · RESTRAINED AWAKENING

The same room warms subtly. No scene replacement.

13.5–15 SEC · HERO HOLD

Stable final composition for approximately 1.5 seconds with dark negative space in the upper third and only minimal breathing/blinking motion.

Platform-optimized production prompt

15s, 9:16 premium stylized animated short. Use the attached image as a strict identity reference; never change character faces, clothes or proportions. Bee is always larger; Match is smaller with the glowing flame chest mark. One continuous intimate nighttime creator studio. Warm amber/charcoal, matte cinematic finish, restrained adult-creator tone.

0–4: Bee at a dark wooden workbench with headphones, compact keyboard, notebook, abstract album-art/storyboard/waveform cards. Thoughtful and dissatisfied, not sad. Slow push. Bee moves one card aside.

4–7: Match is completely hidden before this. Faint gold glow appears behind the final object; Bee notices with eye/antenna reaction, physically moves the object, revealing Match curled asleep. No pop-in.

7–10: Match wakes. Hold a clean two-shot with direct mutual eye contact. Match gives a tiny head tilt, Bee mirrors, flame brightens, Bee smiles. No room transformation yet.

10–12: Match gently floats closer; Bee extends a hand. Both fingertips remain clear in the same frame; touch and hold briefly. Exactly at contact, one small gold pulse originates between fingertips. This must be unmistakable.

12–13.5: The pulse subtly warms the same studio: practical lights, keyboard edge light, one or two art cards, a few particles. No scene replacement.

13.5–15: Stable hero composition hold: Bee larger, Match beside him, both readable, workspace behind, clean dark negative space upper third. Only tiny breathing/blinking.

Camera: slow, controlled, continuous-feeling; no rapid cuts or dramatic zooms.

Audio: minimal ambience, near silence at eye contact, one delicate chime exactly at touch, warm music bloom after. No dialogue or lyrics.

Do not generate: text, logos, UI, portals, new locations, floating music notes, explosions, duplicate/morphed characters, extra limbs/wings/antennae, or bee drawings on the desk.

Priority: reference consistency → hidden reveal → eye-contact hold → fingertip contact → touch-causes-transformation → final 1.5-second hero hold.

9 · Diagnose the correct layer before changing the project

1 · CREATIVE CORRECTION

The idea, message or Voice is wrong. Return to Find Your Flame / Find Your Voice.

2 · INPUT CORRECTION

The reference, source material or context is ambiguous. Fix the input before rewriting everything else.

3 · EXECUTION CORRECTION

The creative brief is right but the platform does not execute it correctly. Change the platform-specific prompt, sequencing or settings.

For The Match, the core creative idea survived. Most of the meaningful problems were input + execution problems, not a failure of the creative direction.

That distinction is one of the bridges into CONTROL / Creator Pro: diagnose drift before changing the wrong layer.

10 · Delivery QA should measure Voice, not just visual quality

VOICE

Does the piece still communicate creative recognition, or did the generation change the meaning?

CHARACTERS

Same faces, proportions, clothing, markings, wing count, antennae count and size relationship?

DISCOVERY

Is Match genuinely discovered rather than already visible?

RECOGNITION

Does eye contact have enough time to register emotionally?

CAUSE + EFFECT

Can the viewer understand that fingertip contact causes the creative awakening?

BRAND DELIVERY

Does the ending support “Create What You Love. Love What You Create.” and leave a stable frame for controlled typography?

11 · Next control layer: CapCut Pro

The production run is not the publishing master. The next step is controlled finishing in CapCut Pro.

LEONARDO PRODUCTION RUN → CAPCUT PRO FINISHING → THREE-PART YOUTUBE SHORTS SERIES → WEBSITE / VIDEO INTEGRATION → ADD YOUR SOUND

The CapCut finishing plan is:

1080×1920 vertical finishing / export master
consistent 1/3, 2/3 and 3/3 series labels
controlled typography rather than generated text
diagnostic editorial overlays where they help teach the lesson
final brand line and URL
audio balance
consistent episode identity
YouTube-ready exports

Those edits are planned, not claimed as completed. We will document the actual decisions after the three finished Shorts exist.

Use the existing CapCut 15-second teaser fundamentals →

Affiliate disclosure: I use CapCut Pro as part of this finishing workflow. I’m a CapCut affiliate, so I may earn a commission if you use my partner link. It does not change what I recommend or what I show in this Dry Run.

CapCut Pro partner link →

12 · The three-Short public case study

THE MATCH 1/3 · FIRST DRY RUN

Purpose: show what worked and what drifted.

SHORT 1 VIDEO — TO BE ADDED

THE MATCH 2/3 · CONTROL THE DRIFT

Purpose: show the reference and instruction improvements.

SHORT 2 VIDEO — TO BE ADDED

THE MATCH 3/3 · PRODUCTION RUN

Purpose: show the optimized generation and controlled final edit.

SHORT 3 VIDEO — TO BE ADDED

All three will be uploaded to YouTube Shorts and embedded back into this training after publication. No placeholder URL is being treated as a live video.

13 · Save the development record for V10

For every serious run, capture:

creator Voice brief
GPT instruction version
reference-image version
production-platform model
dimensions / duration / settings
execution prompt version
what worked
what drifted
what changed next
why the final version was accepted
post-production project and export version

That is how a successful generation becomes a reusable system instead of a lucky result.

Where CONTROL / Creator Pro begins

The free Dry Run gives you the full usable route. CONTROL / Creator Pro goes deeper into reusable GPT instructions, reference architecture, context preservation, controlled iteration, drift diagnosis, cross-tool continuity, version control, reusable prompt systems and delivery standards. Complete Access applies those systems to your actual Voice, Brand, Sound and active work.

Return to the Creator Dry Run Hub →

Current next step: finish and publish the three-part CapCut / YouTube Shorts case study. Add Your Sound follows after the visual production path is documented.

FEATURED IMAGE PENDING
Preferred source: the canonical production reference, an approved workflow visual or the final hero frame after the series package is complete.