Build a Locked Style Block and 5-Slide Prompt Series for HelloFresh
Objective: Write a complete, reusable style block plus 5 individual slide prompts, in the lesson's required prompt order, for a HelloFresh 'Meal Plan Reset' Instagram carousel, so all 5 slides would render as one coherent brand series.
You're a contract content designer for HelloFresh's social team, briefed to produce a 5-slide carousel promoting a new meal-plan reset offer. This is your test assignment before a full retainer.
Freeze a style block first, then write 5 slide prompts in the lesson's required order (canvas and format, style block, layout skeleton in percentages, quoted exact text, one global icon instruction, remaining constraints), changing only the content block between slides.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, and keeps the frozen block copy-pasteable without retyping or accidental rewording
Paid upgrades (optional, faster/deeper)
High visual polish for a hero-quality food-brand carousel once the text and layout are locked
No access? Any image model that accepts a reference image per generation, per the lesson's Tool Routing table
The process
2 steps
Step 01 of 02
The lesson's frozen style block specifies, in order, overall visual style and mood, an exact color palette with a role per color, typography, icon style, and explicit exclusions, then gets pasted verbatim into every generation, never paraphrased.
HelloFresh's brand uses forest green, cream, and a single terracotta accent. What exact style block turns that into a locked, reusable paragraph rather than a vague description?
Procedure
- List the fixed elements first: canvas mood, background treatment, card/surface treatment
- Assign each brand color a hex code and a single named role (never 'accent color' used for two different things)
- State the typography approach and one icon style rule that will cover every icon in the series
- Add explicit exclusions (no photography of raw meat, no drop shadows, no gradients) matching HelloFresh's actual brand constraints
- Freeze the paragraph as final text, this exact wording gets pasted into all 5 slide prompts unchanged
HELLOFRESH STYLE BLOCK (frozen) Canvas: warm cream background, no gradients. Card surfaces are forest green (#2F5233) with 12px rounded corners, no drop shadows. Primary text is cream, secondary text is a muted sage. The single accent color is terracotta (#C0603B), used only for the step-number badge, no other element may use terracotta. Typography: rounded sans-serif throughout, titles bold, body regular, max 8 words per line. Icons: all icons in an identical flat line style, 2px stroke, cream color, no fill. No photography of raw ingredients, no drop shadows anywhere in the composition.
Healthy
The block reads as one locked paragraph with a named role for every color and zero ambiguous terms like 'nice' or 'clean.'
Unhealthy
A vague block like 'earthy, natural, food-brand feel' that gets re-described slightly differently for each of the 5 slides.
What this means
A style block is only useful once it is frozen; if it changes wording between slides, the series drifts even if every individual slide looks fine alone.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Slide 3 uses a slightly different green than slides 1-2 | Paste the frozen block verbatim into every prompt, never re-describe it from memory | 5 min |
Step 02 of 02
The lesson's required order is: format and canvas size, the style block, a layout skeleton in percentage-of-canvas-height zones, the exact text strings in quotation marks, one global icon instruction, then remaining constraints.
Given the frozen style block from Step 1 and 5 pieces of already-finalized post copy, what does a complete, correctly-ordered prompt for slide 3 of 5 look like?
Procedure
- Write the finished post copy for all 5 slides first, before touching the image prompt (Rule: lock copy before generating)
- For each slide, start the prompt with canvas size (1080x1350px vertical) then paste the frozen style block verbatim
- Add a layout skeleton in percentages: 'top 20% headline zone, middle 55% content zone, bottom 25% CTA zone'
- Quote the exact slide copy in quotation marks, never paraphrase it as a description
- Add one global icon instruction and a word-per-line budget, then attach the approved slide 1 image as a reference on slides 2-5
SLIDE 3 OF 5 PROMPT Canvas: 1080x1350px vertical. [frozen HelloFresh style block, pasted verbatim] Layout: top 20% headline zone, middle 55% recipe-card zone, bottom 25% CTA zone. Quoted text: '03 — Swap any recipe in 2 taps.' Single line, max 8 words. Icons: all icons identical flat line style, 2px stroke, cream, no fill. Constraint: no photography, attach the approved slide 1 image as style reference.
Healthy
Every slide's prompt follows the identical section order, with only the quoted text and icon subject changing.
Unhealthy
A prompt that describes the text ('add a headline about swapping recipes') instead of quoting the exact finished copy.
What this means
Consistent prompt order plus quoted exact strings is what makes 5 separately-generated images read as one coherent, on-brand series.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Slide 2's headline text comes back slightly reworded from the approved copy | Quote the exact finished string in quotation marks instead of describing the topic | 5 min |
Final deliverable
One frozen style block (verbatim, reusable) plus 5 complete slide prompts written in the lesson's required order, ready to paste into an image generator.
See a reference example
Flipkart 'Big Billion Prep' carousel — Slide 3 of 5 prompt Canvas: 1080x1350px vertical. [frozen Flipkart style block] Layout: top 20% headline zone, middle 55% content zone, bottom 25% CTA zone. Quoted text: '03 — Track price drops with Price Alerts.' Icons: all icons identical flat line style, 2px stroke, no fill. Constraint: attach the approved slide 1 image as reference, no photography.
Success criteria
You're done when you can:
- The style block is written once, frozen, and pasted verbatim into all 5 slide prompts with no re-wording
- Every slide prompt follows the exact required section order
- All on-image text appears in quotation marks as exact finished copy, never as a description