Practical, tool-agnostic guides for making and refining AI visuals, from Raphael AI.
Status: initial editorial guides, not benchmarked model recipes. The steps below are intended to help structure your own experiments; results depend on the selected model, settings, and references. No specific output or model compatibility is guaranteed.
| Guide | Goal |
|---|---|
| Product visuals | Build a small visual brief, generate concepts, and inspect product fidelity. |
| Character consistency | Establish identity anchors and reduce drift across a short sequence. |
| Prompt iteration | Change one variable at a time and keep a useful experiment record. |
Start creating at raphael.app, or adapt the approach to your preferred tool. For reusable prompt text, see Awesome AI Image Prompts.
Add a focused guide with a clear goal, inputs, steps, checks, and limitations. If you claim a workflow was tested, include the model/version, settings, date, and rights-cleared examples. Follow the community contribution guide.
这里整理 AI 视觉创作的实用工作流,包括产品图、角色一致性和提示词迭代。当前内容为通用操作指南,尚未作为特定模型的效果基准验证。欢迎补充带完整参数和授权信息的实测案例。
Original guides: MIT. Third-party services, models, input assets, and generated outputs may have separate terms.