SEELE TV / Prompt archive

prompt gpt image: model prompt

Use “prompt gpt image” to shape model-specific prompt plan within model-specific prompts. Focus on portable shot contract, dated model-control notes, version and access verification, then keep each variation anchored to the same subject, action, setting, and intended ending.

Production workflowPrompt directions

Search this collection

Collection categories

Generation method

Choose how the production brief begins.

Popular prompts / Model flexible

All video prompts

4 directions

Prompt 01Media reserved

Text generationScene16:915 sec

Prompt Gpt Image — Establish the world

Create a 15 sec 16:9 video direction for “prompt gpt image” in the context of model prompts.
Subject and action: Show the subject entering the setting, performing one clear action, and reaching a visible end state across three connected beats.
Camera and edit: Begin with a restrained wide establishing frame, move to a medium action shot, and finish on one legible detail.
Continuity: Preserve subject identity, wardrobe, hero-object details, geography, lighting direction, and direction of travel across every shot.
Sound: Use original location ambience, tactile action sounds, and no copyrighted music.
Production focus: Make reviewed prompt gpt image workflow and model specific prompts checklist visible through the scene rather than explaining it with text overlays.
Delivery: 16:9, 15 sec, text-led; realistic motion and material behavior; no logos, fake interface, duplicated subjects, impossible physics, or unsupported result claims.

Start Creating Source reserved

Prompt 02Media reserved

Text generationSocial9:1612 sec

Prompt Gpt Image — Build a vertical social cut

Create a 12 sec 9:16 video direction for “prompt gpt image” in the context of model prompts.
Subject and action: Communicate the premise in the first beat, demonstrate one change in the middle, and hold the result long enough to read.
Camera and edit: Use an immediate close opening, two readable handheld reframes, and a stable final composition designed for a vertical crop.
Continuity: Preserve subject identity, wardrobe, hero-object details, geography, lighting direction, and direction of travel across every shot.
Sound: Keep speech original and concise; connect cuts with room tone and motivated practical sounds.
Production focus: Make reviewed prompt gpt image workflow and model specific prompts checklist visible through the scene rather than explaining it with text overlays.
Delivery: 9:16, 12 sec, text-led; realistic motion and material behavior; no logos, fake interface, duplicated subjects, impossible physics, or unsupported result claims.

Start Creating Source reserved

Prompt 03Media reserved

Reference generationReference16:918 sec

Prompt Gpt Image — Preserve a supplied reference

Create a 18 sec 16:9 video direction for “prompt gpt image” in the context of model prompts.
Subject and action: Introduce the referenced subject, show one physically plausible interaction, and return to a composition that can be checked against the source.
Camera and edit: Use the supplied reference as the identity, material, palette, and proportion constraint; cover it with wide, medium, and detail views.
Continuity: Preserve subject identity, wardrobe, hero-object details, geography, lighting direction, and direction of travel across every shot.
Sound: Use grounded ambience and synchronized object sounds; do not imply audio that is not established by the scene.
Production focus: Make reviewed prompt gpt image workflow and model specific prompts checklist visible through the scene rather than explaining it with text overlays.
Delivery: 16:9, 18 sec, reference-led; realistic motion and material behavior; no logos, fake interface, duplicated subjects, impossible physics, or unsupported result claims.

Start Creating Source reserved

Prompt 04Media reserved

Reference generationContinuity4:320 sec

Prompt Gpt Image — Stress-test continuity

Create a 20 sec 4:3 video direction for “prompt gpt image” in the context of model prompts.
Subject and action: Repeat one subject, one hero object, and one spatial relationship through setup, action, reaction, and resolved end frame.
Camera and edit: Stage four connected shots with consistent screen direction, eyelines, lens logic, and motivated camera movement.
Continuity: Preserve subject identity, wardrobe, hero-object details, geography, lighting direction, and direction of travel across every shot.
Sound: Carry one continuous ambient bed across edits and add only original, scene-motivated effects.
Production focus: Make reviewed prompt gpt image workflow and model specific prompts checklist visible through the scene rather than explaining it with text overlays.
Delivery: 4:3, 20 sec, reference-led; realistic motion and material behavior; no logos, fake interface, duplicated subjects, impossible physics, or unsupported result claims.

Start Creating Source reserved

How to use the Model flexible archive

Treat the prompt as a production brief.

Define the model-specific prompt plan behind prompt gpt image

For Prompt Gpt Image, state the subject, environment, action objective, emotional beat, and end condition in plain production language. These invariants keep model-specific prompts variants comparable instead of producing unrelated ideas. Resolve the exact named model or version, then keep creative direction separate from claims about access, syntax, limits, price, audio, duration, quality, or availability. Use only dated first-party evidence for those changing facts. Write the decision owner and intended handoff beside the brief so a search phrase cannot be mistaken for a verified product capability or an instruction that has already been executed. At the next gate, preserve the evidence table, and ask the channel editor to record human approval before the rights check.

Apply the model-specific prompts structure

Write a portable core prompt before adding any vendor or version vocabulary. Keep the subject, action, spatial relationship, shot order, continuity anchors, and end condition independent from a named model. Put model-specific controls in a separate dated note so an unsupported parameter cannot quietly become part of the creative brief. For “prompt gpt image”, keep a visible distinction between required content, optional treatment, exclusions, and facts that need evidence. That structure gives controlled variants a stable baseline and lets a reviewer identify which instruction caused a material change. At the next gate, preserve the review copy, and ask the channel editor to record visible continuity before the rights check.

Prepare authorized inputs for still-image composition

This query calls for an authorized subject reference, intended use, composition, pose or object arrangement, lighting, background, protected details, and delivery format. Specify spatial relationships and light before aesthetic modifiers. Use observable attributes, preserve factual and identity details, and define how the image will be reviewed at its destination size. Record the source, permission, intended audience, protected details, and reviewer before testing language. The goal is a traceable creative proposal, not an assumption that a named platform, model, or workspace will perform the requested action. For this decision, preserve the rights memo, and ask the claims reviewer to record temporal order before the workflow transfer.

Build controlled variants and observable acceptance checks

For this owner, the first review pass should make gpt, image observable rather than merely decorative. Change one meaningful dimension per version and label the hypothesis, invariant details, expected difference, and rejection condition. The brief is reviewable when its creative core remains portable and every model-dependent statement is either cited with a date or clearly labeled as an open test question. Keep failed variants in the decision record so the final wording is explainable rather than selected by impression alone. In the decision log, preserve the rights memo, and ask the brand reviewer to record input provenance before the controlled revision.

Review authorization, safety, and changing product facts

Do not treat a generated suggestion as evidence, a real event, an endorsement, or permission to reproduce another person's identity or copyrighted work. Use only authorized source material, preserve attribution and provenance, avoid deceptive identity or factual claims, and verify any changing product or platform statement in current first-party documentation. Treat every named model, competitor, access term, platform rule, identity use, commercial claim, and output specification as a dated evidence question. If authorization or first-party support is missing, keep the language editorial, stop the handoff, and record the unresolved gap instead of inventing a workaround. For a reversible workflow, preserve the rights memo, and ask the delivery owner to record claim scope before the editorial approval.

Approve a reversible prompt handoff

Confirm the exact provider, version, access surface, accepted inputs, controls, and output behavior against current first-party material before a test. A model name in a query is an evaluation target, not proof that SEELE or any other product currently exposes it. Before approving “prompt gpt image”, check purpose, source rights, identity consent, factual support, continuity, disclosure, destination requirements, and the named review owner. Preserve the original brief and version notes. This handoff supplies authored guidance only and neither runs generation nor guarantees access, quality, speed, licensing, publishing, downloads, or outcomes. For a reversible workflow, preserve the reference set, and ask the delivery owner to record failure conditions before the editorial approval.

Evidence boundary

Use the source, rights and current workspace controls as the acceptance boundary.

Before you hand off

Questions to resolve.

Does this “prompt gpt image” page execute a prompt or generate media?

No. It is an editorial planning and evaluation page. It does not upload a source, call a model, expose verified controls, execute generation, publish media, or display a generated result. For this decision, preserve the failure note, and ask the rights reviewer to record visible continuity before the evidence refresh.

How should a reviewer evaluate this model-specific prompt plan?

Keep one stable brief, change one instruction at a time, record the expected difference, and compare the result with explicit acceptance and rejection criteria. The brief is reviewable when its creative core remains portable and every model-dependent statement is either cited with a date or clearly labeled as an open test question. For this decision, preserve the authorization record, and ask the workflow owner to record input provenance before the evidence refresh.

What evidence boundary applies to these model-specific prompts?

Use only authorized source material, preserve attribution and provenance, avoid deceptive identity or factual claims, and verify any changing product or platform statement in current first-party documentation. Verify changing model, product, price, access, policy, license, input, output, and entitlement statements in dated first-party sources. Search demand and prompt wording are not capability evidence. For this decision, preserve the rights memo, and ask the production lead to record reversal cost before the delivery pass.

Continue the workflow

Take a prepared brief into the workspace.

Continue in the SEELE workspace to inspect the currently available Film & CG workflow.

Try it free