正在加载视频...

视频加载失败

Your agent within Google Flow can be a helpful creative partner, guiding you from concept brainstorming to final creation. For example, your agent can help you explore storyboard directions based on your idea. This is great for both planning out key scenes and locking in details. Once you finalize...

54,948 次观看 • 9 天前 •via X (Twitter)

23 条评论

Ferdinand 的头像
Ferdinand9 天前

The agent is super dumb, I wonder which model powers it. Secondly every vid gen looks so polished (obvious AI look) zero prompt adherence often. always rejecting some prompts (aka violation) fix ts, drop better models, quit with this flash bullshit

Chinez 的头像
Chinez9 天前

@FlowbyGoogle @joshwoodward remove censorship too, it stifles creativity.

Vandalist 的头像
Vandalist9 天前

Creating an avatar always failed on Flow, how to solve this please

Dexter Makachikadev 的头像
Dexter Makachikadev9 天前

Fix the moderation, I can't make anything other than cartoons. I had to use seedance for other stuffs.

puzzleheadrise5099 的头像
puzzleheadrise50999 天前

Always rejecting the prompts You have horrendous censorship

FengDaQuant 的头像
FengDaQuant9 天前

Y'all need to change your censorship. I can't generate videos because my prompt gets rejected claiming that I'm using prominent figures faces but that's not true. I'm using characters, I've created from scratch.

anish 的头像
anish9 天前

Idk man melius does all of that and better

Raja 的头像
Raja9 天前

Increase 10 second limit.

Isoldegwow 的头像
Isoldegwow9 天前

Storyboard, lock it in, agent makes the final video — so my job is basically approving things and taking credit.

Tashford AI 🇪🇬 的头像
Tashford AI 🇪🇬9 天前

Even the official video from Google AI is giving up Please update your model It's sensitive to Prompts

Sarah Parker 的头像
Sarah Parker9 天前

Lock the frames first. Then let the video keep the face.

Momo D NH 的头像
Momo D NH9 天前

STOP Watermark as a 4 Millions Influencer!

AI & Space News 的头像
AI & Space News9 天前

The jump from brainstorming to a finished video in one workflow is pretty exciting. Visual consistency will be the real test.

RamaXwhale | (❖,❖) 的头像
RamaXwhale | (❖,❖)9 天前

your agent needs to be improved that's all

mohammed salman 的头像
mohammed salman9 天前

Pls bring back low priority model

Tomasz Proshowski 的头像
Tomasz Proshowski9 天前

Dear Google thank you very much for more than amazing news.👏👏👏👏👏👏👏👏👏👏👏👏

مــحمــد | MK 的头像
مــحمــد | MK9 天前

Please we need the videos 1080 DPI, not just upscale! I think google flow if didi this! It will be the best AI platform

Benjamin | Meta Ads Video Editor 的头像
Benjamin | Meta Ads Video Editor8 天前

Agent is a terrible tool icl in flow , the model isn’t that smart

anaknijihoz.ron 的头像
anaknijihoz.ron9 天前

ohhh interesting gonna try this

真田広之 的头像
真田広之9 天前

It is hoped that a set of prompt words can generate two videos, each lasting 10 seconds, and the two videos are consecutive.

Henrique Dalposso 的头像
Henrique Dalposso9 天前

agent is dumb af

Chinez 的头像
Chinez8 天前

@alisa_fortin please eliminate censorship from Omni Flash, it stifles creativity.

Hershal Rao 的头像
Hershal Rao9 天前

great, now my ai agent can reject my storyboard ideas too

相关视频