Fawna Illustrate
Illustrate is text-to-image. Clean detail, strong composition, fast turnaround. Use it for hero stills, product shots, mood boards, and anywhere you need a single beautiful frame from a text prompt.
- Tier label
- Illustrate
- Price
- From 6 credits per image
- Aspect ratios
- 1:1, 4:5, 3:4, 16:9, 9:16
- Outputs per generation
- 1, 2, or 4
- Quality tiers
- Fast, Standard, Ultra
- References
- None (text-only)
- Keyframes
- Not applicable (still image)
- Negative prompt
- Supported
- Magic Prompt
- Supported
When to pick Illustrate
- You need a still image, not a moving one.
- You do not have reference photos and prefer to describe what you want in text.
- You are generating thumbnails, mood boards, or hero shots.
- You want the cheapest way to generate high-quality stills.
Quality tiers
Illustrate exposes three quality tiers. Same prompt, different time and credit budget.
| Tier | Use for | Relative cost |
|---|---|---|
| Fast | Ideation, first-pass drafts, thumbnails | ~0.5x baseline |
| Standard | Default for most work | 1x baseline |
| Ultra | Hero stills, print-quality output | ~3x baseline |
Prompt recipe
Start with "A photo of..." for photoreal. Start with "An illustration of..." or "An oil painting of..." for stylized.
A photo of a tangerine resting on a folded linen napkin beside a chipped enamel plate. Soft overhead morning window light casting long shadows. Shallow depth of field, 50mm macro lens, natural muted palette, subtle film grain, minimalist still life composition.
An ink and watercolor illustration of a hot-air balloon drifting over a patchwork of farmland at sunrise. Soft wet-on-wet edges, warm earth palette, thin ink linework, visible paper grain, children's book aesthetic.
Outputs per generation
Pick 1, 2, or 4 outputs. Each is an independent attempt at the same prompt. Four outputs per roll is usually the right choice because the cost is per-image and the variation lets you pick the strongest frame.
Aspect ratios
Illustrate composes differently for each aspect ratio. A 16:9 version of the same prompt will frame wider context; a 1:1 will center the subject more tightly. If you need a specific use case (mobile story, blog header, square thumbnail), pick that aspect from the start rather than cropping after.
Limits
- No reference input. If you need to keep a face consistent across images, use Compose.
- No animation. For moving output, use a video tier.
- Text rendering inside images is imperfect. Expect a 50/50 chance of clean spelling for 3+ word strings.
Tips
- Prefix your subject with "A photo of..." for realism or a medium word for style.
- Run 4 outputs at Standard first, pick the winner, then re-roll on Ultra for the final.
- Skip the word "4K". It biases the model toward oversaturated, plasticky output.
Where to go next
- Fawna Compose for reference-driven image gen.
- Writing a great image prompt for the full prompt playbook.