以最少提示詞轉換來創作圖像的 AI 代理。
這是一款由 LLM 驅動的通用圖像創作代理。不同於會將提示詞直接傳送給所選圖像模型的 Image Generator,Image Agent 能理解你的需求、記住對話內容,並協助你使用不同的圖像模型,同時將提示詞轉換降至最低。
從英文翻譯,僅供參考。
I’d like to create a promotional image showcasing San Francisco’s scenery. What text-to-image prompt would best capture the city’s distinctive character? Use that prompt to generate an image for me.
Prompt used:
Create a premium travel-promotion photograph of San Francisco in a wide cinematic composition, viewed from the Marin Headlands. Make the International Orange Golden Gate Bridge the hero, spanning the blue-green bay, with wisps of Pacific fog curling beneath it and San Francisco’s rolling hills and skyline in the distance. Warm late-afternoon sunlight breaks through cool coastal haze, illuminating the bridge, golden grasses, and coastal wildflowers. Capture the city’s distinctive blend of dramatic coastal nature, iconic engineering, urban energy, and dreamy fog. Photorealistic editorial landscape photography, rich but natural colors, crisp architectural details, atmospheric depth, and an inviting, aspirational mood. Leave uncluttered sky in the upper left for a future headline. Landscape 3:2 format. No text, logos, watermarks, or collage.
The generated image combines the bridge, fog, skyline, and golden coastal scenery in a 1536 × 1024 composition. I checked the result visually; it leaves clear sky for promotional copy.
本次執行費用85 點數≈ $0.085圖片生成 ×1GPT-6.1 Sol · 41,160 詞元