从文字创造新图片，或用参考照片探索新风格。使用免费积分体验当前可用的图片模型。
Eight copyable text to image briefs tuned for Dola AI: concrete subject, composition, material, lighting, and ratio.
Premium brand-world image for an imaginary fragrance called Obsidian Bloom, black volcanic glass bottle on wet basalt, white jasmine petals, thin red thread, misty dawn light, restrained luxury campaign still, crisp product silhouette, no logo, no watermark.
Museum campaign still of a single walnut chair floating inside a quiet concrete gallery, long soft shadow, dust in skylight beams, minimal composition, refined architecture photography, wide editorial composition, no readable text, no logo, no watermark.
Three-quarter product study of a handmade ceramic smart speaker with tiny glaze imperfections, placed on raw linen beside a brass volume dial, soft window light, gentle shadow gradient, calm premium product photography, no text, no logo, no watermark.
Wide cinematic retail window scene of a tiny late-night bakery glowing in rain, trays of pastries, condensation on glass, one warm pendant lamp, reflections of street lights, cozy but sophisticated commercial photography, no text, no logo, no watermark.
Cute but polished 3D mascot character, a moon courier wearing a tiny padded messenger jacket and carrying a crescent-shaped satchel, clean studio background, soft vinyl texture, expressive eyes, product mascot quality, no text, no logo, no watermark.
Editorial collage still of old contact sheets, torn blueprints, pressed leaves, a silver fountain pen, and one translucent acetate overlay on a mahogany desk, top-down composition, warm lamp light, tactile archival mood, no readable text, no logo, no watermark.
Editorial food image of tangerine segments arranged like a precise flower on a handmade white plate, pale blue tablecloth, tiny salt crystals, silver fork, crisp natural window light, magazine-quality restraint, no text, no logo, no watermark.
Miniature harbor diorama inside an open travel notebook, tiny paper boats, folded blue paper waves, brass compass, soft morning light, shallow depth of field, tactile craft meets realistic photography, clean wide composition, no text, no logo, no watermark.
Pick the image model by the task: source photo control, readable text, high-fidelity stills, fast concepts, or precise local changes.
Use GPT Image 2.5 for detailed creative briefs, reference-guided product edits, structured layouts, and polished scene generation with the connected high-quality preset.
Choose the available GPT Image 2.5 settings and review the full credit quote before generating.
Bring an approved image into a new setting. Describe the change and the product, subject or layout details that should stay unchanged.
Available resolution, aspect ratio and quality controls depend on your selected model. Review fine text and edges in the downloaded result.
Create AI images from prompts, or upload a source photo when you want an AI image edit. Compare Nano Banana, Nano Banana Pro, and GPT Image 2 in one studio.
Use Nano Banana Lite or GPT Image 2 for new prompt images, then switch to image to image workflows when an upload should guide the edit.
Use a source image to guide restyling, cleanup, local edits, or controlled variants while keeping the useful structure of the original photo.
Use AI image generation for fresh concepts, product stills, edits, and controlled variants in one workspace.
Create campaign stills with Nano Banana Pro or GPT Image 2 when polish and prompt-following matter.
Use Nano Banana for product restyling and GPT Image 2 for detailed edits before assets reach catalogs or ads.
Develop branded imagery, poster directions, and cohesive visual reference systems with Nano Banana Pro or GPT Image 2.
Use GPT Image 2 or Nano Banana to adapt an existing image while preserving the useful source structure.
从一个清晰的想法开始，提交前检查设置与积分费用。
写明主体、环境、光线与风格；使用参考图片时说明需要的变化，视频则补充动作与镜头运动。
选择可用模型及支持的格式。登录账号后，在生成按钮上确认当前批次的积分费用。
确认设置后点击生成，在编辑器中查看任务状态。完成时间随模型、设置及服务负载变化。
检查结果，修改提示词继续尝试。已完成的作品可在结果区或登录后的历史记录中查看和下载。
了解免费额度、账号要求与图片和视频创作流程。
可以，前提是所选图片模型支持参考图。上传原图后描述想要改变的场景、风格或构图。参考数量和输出设置以当前模型支持的选项为准。
新用户可获得 100 注册积分，每日签到可领取 40–120 积分。签到奖励按连续签到天数计算，每个 UTC 日可领取一次。生成会消耗积分，免费积分不代表无限生成。
可以先浏览示例、写提示词并选择设置。生成或上传参考素材前，需要登录账号。登录成功后仍由你点击生成。
费用取决于模型、分辨率、质量、视频时长及结果数量。生成按钮会显示当前批次的积分费用，请在提交前确认。余额不足时可以领取可用签到奖励，或查看定价。
完成的作品会显示在结果区，也可以在登录账号的历史记录中查看。先预览图片或播放视频，再下载需要保留的结果。
发布前请阅读本站条款与所选模型的使用规则，并确认你有权使用上传的图片和其他素材。不要仅凭生成成功判断所有使用场景都已获授权。
从提示词或参考图开始，按当前可用模型和积分余额安排你的创作。
