Black Forest Labs
How to Create AI Ad Creative with FLUX.2
Generate the visual layer of an ad-creative matrix on Astorie using FLUX.2 — lifestyle hero photography, product context shots, and aspirational scenes that feed downstream Ideogram nodes for headline overlay. FLUX.2 is the prompt-fidelity workhorse: it renders specific products, environments, and compositions almost literally, which is exactly what performance marketers need when the creative brief reads like a shot list. For text overlays, chain to Ideogram; for the photography itself, FLUX.2 is the cleaner pick.
Step-by-Step Guide
Write the brief as a shot list, not as a vibe
FLUX.2 reads literal compositional prompts. Skip "energetic gym vibe" and write: "A runner mid-stride on a coastal trail at sunset, warm golden side light from camera right, athletic shoes prominently visible in lower third, shallow depth of field, 1:1 aspect." The shot list gives FLUX.2 the structure it needs; vibe words leave it guessing.
Pin the product position with explicit language
For ad work the product must be in-frame, in-focus, and in a known position. Write "athletic shoes prominently visible in lower third, focal point, sharp focus" or "moisturizer bottle held in right hand, label facing camera, mid-frame." FLUX.2 follows literal positions — vague product placement is the most common ad-creative failure mode.
Generate the photo without text, then chain to Ideogram
FLUX.2 renders typography but Ideogram wins on legibility for production ads. The cleanest pipeline: FLUX.2 generates the photographic layer (lifestyle, product, scene), then a downstream Ideogram V3 node takes the FLUX.2 output as a reference and adds the headline + CTA in legible bold typography. Two nodes, one canvas, ad-ready output.
Fan one photo concept into every aspect ratio
Duplicate the FLUX.2 node four times, set 1:1 (Meta feed), 4:5 (Meta vertical), 9:16 (TikTok/Reels), 16:9 (YouTube preroll). Keep the prompt identical — only the aspect changes. FLUX.2 recomposes per ratio while holding the photographic voice. Each output then feeds its own Ideogram node for ratio-appropriate headline placement.
Iterate one element at a time across A/B variants
For variant testing, change ONE variable per duplicate node — light direction, time of day, palette accent, model demographic, prop. Keep everything else identical. FLUX.2 produces near-identical compositions with only the changed element shifted, which is exactly what performance marketing needs for clean A/B reads.
Save the canvas as an ad template per concept
Once a winning ad concept is locked, save the canvas as an Astorie template — FLUX.2 photo node, downstream Ideogram nodes, aspect-ratio fan-out, parameter pins. Next campaign, swap the subject prompt and the headline string. The ad pipeline becomes reusable: same template, different campaign, identical production cadence.
Prompt Examples
Hero photo for Meta feed. The "no text overlays" instruction keeps FLUX.2 focused on photography — text gets added downstream by Ideogram.
A runner mid-stride on a coastal trail at sunset, warm golden side light from camera right, athletic shoes prominently visible in lower third with sharp focus, shallow depth of field on the background, magazine-quality crop, 1:1 aspect, no text overlays
TikTok vertical that bakes copy room into the composition. The right-third offset leaves space for the downstream Ideogram headline.
Same runner concept, vertical composition with the figure offset to the right third for caption space on the left, 9:16 aspect, no text overlays
Product-in-context lifestyle. The literal product placement language ("bottle held in her right hand, label facing camera") is the load-bearing prompt move.
Lifestyle aspirational scene: a woman applying moisturizer at a sunlit bathroom mirror, soft natural light from a window on camera left, the moisturizer bottle held in her right hand with the label facing camera, sharp focus on the bottle and her face, soft focus background, 4:5 aspect, no text overlays
A/B variant. Single-variable change keeps the test clean — only light and palette differ between A and B.
Same lifestyle scene with one variable changed: late afternoon golden light instead of morning daylight, warm peach palette instead of soft blue, otherwise identical composition, 4:5 aspect
Parameter Tips
Pin FLUX.2 Pro for hero ads, base for variant exploration. Pro's prompt adherence is meaningfully better on detailed product placement.
Always include "no text overlays" in FLUX.2 ad prompts — it stops the model from rendering placeholder text and keeps the photographic layer clean for downstream Ideogram overlay.
Use literal product placement language ("lower third, sharp focus, label facing camera"). Vague placement is the top failure mode for ad creative.
For lifestyle scenes with people, name the demographic, action, and product interaction explicitly. Generic "person using product" returns generic person.
Bake copy room into compositions destined for Ideogram overlay — offset the subject to one third, leave the opposite third clean for headline placement.
Match aspect ratio to the placement (1:1 Meta feed, 4:5 Meta vertical, 9:16 TikTok/Reels, 16:9 YouTube preroll). FLUX.2 recomposes per ratio.
What to Expect
FLUX.2 returns 1024-2048 wide photographic outputs that hold composition and product placement across ratio variants from one prompt. Generation time: base 15-30s, Pro 30-60s. Output is the photographic layer for ads, not the finished ad — chain into Ideogram for headline overlay before NLE export. The FLUX.2 + Ideogram two-node pipeline is the cleanest production workflow for ad creative on Astorie.
Use FLUX.2 on Astorie
Connect FLUX.2 with other AI models on Astorie's infinite canvas. No GPU required — start free.
Get Started FreeRelated features
Docs
Related reading
Try Other Models for This Task
Ideogram
Ideogram V3
Fan one ad concept into a 30-asset paid-social matrix on Astorie using Ideogram V3 — every aspect ratio, every CTA variant, every headline localization, with the copy rendered legibly inside the image. Ideogram is the only AI image model that handles in-image text reliably; for performance-marketing work where the headline IS the asset, no other model in the stack comes close. Pair it with a saved brand prompt, fan into ratio-specific nodes, and ship a Meta + TikTok + YouTube test matrix without a Figma round-trip.
View guideOpenAI
GPT Image 2
Generate ad creatives on Astorie using GPT Image 2 — strong reasoning over compositional briefs and multi-element prompts, with reasonable in-image text rendering for short headlines and CTAs. GPT Image 2 sits between Ideogram (typography champ) and FLUX.2 (photographic prompt fidelity): it understands narrative briefs better than either, which makes it the right pick for ad concepts that require interpretation ("show a busy parent finding 5 minutes for themselves") rather than literal staging. Pair it with a downstream Ideogram node for the final headline pass when typography needs to be perfect.
View guide