I want a 60-second vertical recipe video for fried rice

## Output structure

<instructions>
Create the complete 60-second vertical fried-rice video package itself, not a table of contents. Begin with a cut table containing cut number, visual, camera, audio, and duration in real seconds; make all durations total exactly 60 seconds. Follow it with one copy-paste-ready generation prompt per cut, a single noun-phrase exclusions line, parameter values, and concise handoff guidance. Separate engine-ready text from user-facing explanations visibly. Include one line identifying any staging change and why it improves readability.
</instructions>
<context>
The subject is fried rice. The user supplied no ingredients, quantities, cookware, setting, presenter description, visual style, or audience. Preserve the 60-second vertical format and do not invent those facts.
</context>
<output_format>
Use a hybrid format: tables and compact bullet lists for production values; short prose only for rationale, continuity, and handoff. Do not label the internal sections with tool names.
</output_format>

## Render direction

<instructions>
Rebuild the material as directed food video rather than transcribing it. Choose the number of cuts from the 60-second duration: allow roughly two seconds for a readable static beat and three or more seconds when an action must be understood. Allocate time unevenly, giving the densest coverage to the decisive cooking transformation and enough time for the final plated result.

Engine-ready prompts must use comma-joined noun phrases in this order: subject, action or state, composition and framing, lighting, material and texture, visual style. Show physical causation in sequence: ingredients enter the pan, heat changes their state, the utensil moves them, and the finished rice is plated. Describe light by its visible contact with food, cookware, hands, and counter surfaces.

Settle the unnamed production constants in cut 1 and repeat them verbatim thereafter: a clean neutral countertop, one consistent overhead three-quarter camera angle, soft daylight from frame left, natural food colour, and the same cookware and utensil throughout. Keep the chosen visual treatment fixed across cuts. Preserve the user's facts; where a fact is absent, mark it for input rather than fabricate it.
</instructions>
<context>
The user wants a 60-second vertical recipe video for fried rice. No subject anchor or additional visual constraints were supplied.
</context>
<output_format>
Provide the noun-phrase prompts inside clearly marked engine-ready blocks. Put explanations and the one-line staging-change rationale outside those blocks in readable prose.
</output_format>

## Exclusions

<instructions>
Use one continuous noun-phrase negative prompt. Exclude only concrete defects that would undermine a readable recipe video: distorted hands, extra fingers, duplicate utensils, changing cookware, changing countertop, inconsistent food colour, unreadable ingredient quantities, obscured pan, blocked action, abrupt continuity changes, watermark, logo, subtitles, unrelated ingredients, unsafe raw-food handling, burnt food, artificial steam.
</instructions>
<context>
The requested output is a cooking video, so exclusions must protect hand action, ingredient continuity, food readability, and clean presentation without adding unsupported stylistic restrictions.
</context>
<output_format>
Write the exclusions as one comma-separated line, not as sentences.
</output_format>

## Continuity procedure

<instructions>
This is a multi-cut sequence. Record the first generated result's seed and reuse it for every later cut if the service supports a fixed seed. If it supports a reference image or reference frame, use cut 1 as the reference for subsequent cuts. Wording alone will not reliably preserve identity or set continuity: reference input is stronger than a fixed seed, and a fixed seed is stronger than text description; a seed remains useful only while the prompt remains unchanged.

In cut 1, settle and state the neutral countertop, cookware, utensil, camera angle, frame-left daylight, colour treatment, and any presenter or ingredient details that become available. Quote that continuity sentence verbatim in every later cut. Keep the fixed points unchanged while varying only the cooking action, ingredient state, framing distance, and timing. Match the final plated composition to the opening composition if a seamless loop is requested; otherwise do not assume a loop.
</instructions>
<context>
No presenter, cookware, ingredient list, or loop requirement was provided. Do not create a subject anchor. Any newly settled visual constants must remain stable across all cuts.
</context>
<output_format>
Place the verbatim continuity sentence in a dedicated continuity line inside each per-cut prompt, then explain the seed/reference procedure in user-facing handoff prose.
</output_format>

## Parameters and handoff

<instructions>
List these values together: aspect ratio `9:16` as a vertical frame suited to short-form viewing; resolution `[RECOMMENDED: 1080p — sufficient detail for readable food preparation in a vertical social-video frame; change it if the destination requires another size]`; frame rate `[RECOMMENDED: 30 fps — natural motion for hand and utensil actions; change it if the intended platform requires another rate]`; total length `60 seconds`; audio `[FILL IN: narration, music, cooking sounds, or a combination]`.

Keep the recipe factual. Do not invent ingredients, quantities, cooking times, temperatures, cookware, or safety claims. If ingredient or process facts are unavailable, identify them as open inputs instead of filling them with plausible details. The parameters must contain values only, with no tool names, slash commands, or double-dash flags. Add: “The service plan may not offer the chosen resolution; step down one level if unavailable.”</instructions>
<context>
The fixed user facts are a 60-second duration, vertical format, and fried rice subject. Ingredients, quantities, equipment, setting, and audio direction remain unknown.
</context>
<output_format>
After the parameter block, number only the remaining fill-in items:
1. `[FILL IN: ingredients and quantities]` — supply the exact ingredients and amounts used.
2. `[FILL IN: cooking setting and available cookware]` — state the location, pan or wok, heat source, and utensil.
3. `[FILL IN: audio direction]` — choose narration, music, cooking sounds, or a combination.
4. `[FILL IN: preferred visual style or intended audience]` — state the desired look or who should watch.

Then offer: “(1) Fill these in and I will revise the video package, (2) keep the recommended values and finish it as it stands, or (3) change the structure first.” Wait for the user's choice.
</output_format>