AI Agent Poster Brought to Life

MiniMax H3WorkflowCreativeImage to VideoCat10s
Turning a Godfather-style poster into a 10-second animated video: text fading in, a Hitchcock zoom, a freeze on the cat's eye. Animating a poster is a genuinely useful little workflow — one prompt is all it takes.
PROMPT · Video Prompt
Create with AI
[Workflow Prompt] "Use the Minimax H3 model to generate a 10-second 768P video based on this reference image, with the following prompt: the camera pushes in, the camera slowly circles around, the text 'The Godfather' at the top slowly transforms into 'The PixelAIGC' as that hand adjusts it, the character is revealed behind the text using a Hitchcock zoom (dolly zoom), and it finally freezes on the cat's eyes held in the hand and the text above." About 6 minutes later, MiniMax H3 created an extremely detailed process and generated a video — the result was excellent, and I only tried it once. I'm sharing a screenshot of my conversation in the comments. #MiniMaxH3 on #MiniMaxDesign 0:03 — AI-generated. Paid partnership.
✍️ Editor’s Notes

这条本质是一段带炫耀口吻的工作流笔记,真正的生成指令被压缩成一句话(镜头推进+环绕+文字变形+希区柯克变焦+定格猫眼),信息密度很低,需要补全才能直接拿去跑图,这也是被标need_expanded的原因。MiniMax H3处理「静态海报转10秒动态」这类任务时,「希区柯克变焦」这种业内黑话式镜头语言反而是最值钱的一个词,值得单独拎出来放大写清楚触发时机。

Want tighter control? Editor-expanded reference version (not the original prompt):
参考海报静态图,生成10秒768P视频。镜头缓慢推进并同步做小幅度环绕运动,机位始终对准画面主体。海报上方文字「The Godfather」随着画面中那只手完成一次自然调整动作后,逐字渐变、笔画重组为「The PixelAIGC」,字体质感与原版保持一致,无闪烁跳变。人物主体在文字变化的同时使用希区柯克式变焦(推轨反向变焦/Dolly Zoom),背景透视被拉伸挤压出眩晕感,人物本身大小基本不变,逐渐显现在文字后方,形成前景文字、后景人物的双层构图。全程环境光维持海报原有的冷暗色调,仅在人物显现瞬间给一次极轻微的高光提亮。结尾镜头快速收拢,最终定格在人物怀中猫的双眼特写,猫瞳孔清晰锐利,与画面上方仍然可见的「The PixelAIGC」文字同框收尾,无额外转场,无多余文字浮现。
Create with AI
工作流提示词
PROMPT · Full Prompt
Create with AI
"Use the Minimax H3 model to generate a 10-second 768P video based on this reference image, with the following prompt: the camera pushes in, the camera slowly circles around, the text 'The Godfather' at the top slowly transforms into 'The PixelAIGC' as that hand adjusts it, the character is revealed behind the text using a Hitchcock zoom (dolly zoom), and it finally freezes on the cat's eyes held in the hand and the text above." About 6 minutes later, MiniMax H3 created an extremely detailed process and generated a video — the result was excellent, and I only tried it once. I'm sharing a screenshot of my conversation in the comments. #MiniMaxH3 on #MiniMaxDesign 0:03 — AI-generated. Paid partnership.
✍️ Editor’s Notes

和上一条几乎是同一份素材的重复贴出,唯一区别是少了开头那句工作流总述,直接从引号里的核心指令开始——这恰好说明这批「分享截图」类帖子里,真正有效的prompt往往就藏在被引号包起来的那一小段,前后的感叹和评论都是噪音,摘取时得学会剥离。

Want tighter control? Editor-expanded reference version (not the original prompt):
参考海报静态图,生成10秒768P视频,镜头持续缓慢推进并小幅环绕,始终对准画面主体不偏移。画面上方文字「The Godfather」在那只手完成一次轻微调整动作的过程中缓慢过渡、重组为「The PixelAIGC」,笔画变化自然连贯,不出现闪烁或突变。人物按希区柯克变焦手法呈现:镜头推近的同时反向拉远焦距,背景空间被拉伸出明显的眩晕纵深感,人物比例保持稳定,逐渐从文字后方浮现,形成清晰的前后层次。色调延续海报原本的冷暗基调,人物完全显现时给予一次克制的高光过渡。视频最后收束在人物怀中猫咪的双眼特写上,眼神清澈锐利,与画面上方仍留存的文字同时定格作为结尾画面,不加多余特效。
Create with AI
Related Works
×