Generate illustrations and marketing images with AI
Skills used in this guide (3)
用 fal.ai MCP 统一生成图像、视频与音频：文生图、图/文生视频、语音合成、视频配音都有对应模型。适合一句话生成图片视频音频的场景。
npx skills add affaan-m/ECC -g用 Stability AI（SD3.5、Ultra、Core）生成图像：文生图、图生图、修复、放大、去背景、搜索替换，15 种艺术风格。
npx skills add sickn33/agentic-awesome-skills -g（占位）源数据描述为空
npx skills add nexu-io/open-design -g
The worst text-to-image is "one-line prompt, pure luck". The point: set style ref + subject + composition + ratio as structured prompt, then iterate small — not throw "draw something nice" and hope.
Who this is for / not for
For:
- Marketing / article / social images
- Have a direction but can't write prompts
- Batch same-style images
Not for:
- Specific real person portrait (rights/ethics)
- Pixel-exact (text-to-image isn't precise)
- Brand-logo-level retouch (that's design)
Before you start
- Use & ratio: cover 16:9 / square 1:1 / portrait 9:16.
- Style ref: 1-2 refs or description.
- Subject & mood: what, what feeling.
Steps
1. Install the skill
fal-ai-media is a media-gen skill. Copy its folder into your client's skills directory (same as any Agent Skill):
| Client | Directory |
|---|---|
| Claude Code | ~/.claude/skills/ |
| Codex | ~/.codex/skills/ |
| Cursor | ~/.cursor/skills/ |
| Trae | ~/.trae/skills/ |
| CodeBuddy | ~/.codebuddy/skills/ |
| Tongyi Lingma | ~/.lingma/skills/ |
2. Structured prompt
SaaS product newsletter cover, 16:9:
subject = laptop with floating charts, style = flat vector + soft gradient + soft shadow,
palette = blue-purple, mood = professional yet light. Give prompt first, then image.
No text watermark, no real face.3. Switch engine for style
SD family → stability-ai; stylized → venice-image-generate, mood-first.
4. Iterate small
Not happy with v1? Point the specific fix ("bg too busy→whitespace", "subject small→bigger"), don't rewrite all.
Pitfalls
| Pit | Symptom | Fix |
|---|---|---|
| One-line prompt | Pure luck | style+subject+comp+ratio structured |
| Wrong ratio | Distort | Fix ratio upfront |
| Real face | Rights/ethics | Use illustration/silhouette |
| Text watermark | Unusable | Prompt "no text" |
| One-shot | Wasted imgs | Small iterative fixes |
FAQ
Write prompts?
Style ref + subject + composition + ratio + mood; structured beats adjectives.
Set ratio?
Yes — cover 16:9, square 1:1, portrait 9:16, fix first.
Which engine?
fal general, SD controllable, venice stylized; pick by use.
Batch same style?
Yes — fix prompt template, swap subject only, consistent.
Rights risk?
Avoid real faces & copyrighted; confirm engine license for commercial.
Related guides
Continue across these scenes:
相关源码
- affaan-m/ECC · fal-ai-media source
- sickn33/agentic-awesome-skills · stability-ai source
- nexu-io/open-design · venice-image-generate source
- Agent Skill open format spec
一句话结论
结论：文生图用结构化提示词（风格+主体+构图+比例）小步迭代，别扔一句"画个好看的"。
References
Related Scenes
Generate icons and article illustrations with AI
The pain of illustration isn't "can't draw" — it's "inconsistent style, blurry when scaled, can't edit". The point of AI vector illustration: set style and size first, want vector not bitmap, so it scales sharp and the source stays editable.
Make podcast scripts and voiceover with AI
The pain of podcast/talking-head isn't "no voice" — it's "script reads like a paper, voice sounds robotic, nobody finishes". The point of AI: write the script as spoken human language, pick a natural voice, then decide on visuals — don't pile effects upfront.
Brainstorm and mind-map with AI
The worst brainstorming is "diverge and stop". The point of AI help: first force enough candidates (quantity before quality), then converge on 3 doable ones by "feasibility / impact" — not stop at "many ideas".
Turn the architecture in your head into a diagram with AI
The pain of diagramming isn't "can't draw" — it's that edits are exhausting: change one box and you re-layout the whole thing. Generating descriptive, re-flowable diagrams turns "drawing" into "describing structure"; editing text is 10x faster than editing the picture.