
👨💻 前端工程师 & 出海探索者 🚀 英推增长 (DM 领 7 折券) - https://xitter.gq/t.co/2LyZXXb5H2 📢 分享我的 SaaS 产品出海之旅 | 英文推: @devinzhang_me
Joined August 2022
- Tweets1.5K
- Following612
- Followers3.6K
- Likes3.2K
有人说这是我请花钱找人制作的,也有人说这是我请朋友帮忙录制的,甚至还有人以为这就是我,离了大谱 😅
实际上这是 AI 数字人,可以使用这个工具根据链接生成宣传视频(免费额度是有水印的)
video.creatify.ai/link2video…
seedance 2.0 is getting scary realistic
this doesn’t look like ai ad output anymore
it looks like content pulled from someone’s camera roll
product table shot
creator holding the bottle
random lifestyle scene
messy organic food visual
that’s the edge
the ad doesn’t feel like a polished ad
it feels like proof that already existed
for beauty, supplements, wellness, and food brands this is huge
you can now generate the “organic filler” creatives that used to require creators, shipping, reshoots, and weeks of waiting
same product
different environments
different proof angles
different reasons to believe
i put the full realistic ugc prompt workflow into one file
rt + comment "realugc" and i’ll send it
(follow for dm)
阿飞 retweeted
Forget Sora. Kling 3.0’s logic is the new gold standard.
This model has a precise logic unlike anything else on the market right now.
I spent +$100 testing its limits so you don’t have to.
RT + comment "Guide" and I’ll DM you my full breakdown of the multi-prompt logic.
BREAKING: I found a way to make AI videos that don’t sound like AI
sora 2, veo 3.1 are great but they all share the same issue:
that robotic AI accent
the cadence everyone recognizes instantly
so people swap in elevenlabs
and now it sounds like:
- a $5,000 mic
- a soundproof studio
- a podcast intro
but that’s not UGC
if the video looks like it was filmed in a bedroom
the audio shouldn’t sound studio-perfect
it should sound like it’s actually in the room
the workflow we use removes the fake AI voice feel
- no robotic cadence
- no over-polished voices
- no obvious “this is AI” giveaway
why this matters:
when voices sound too clean or robotic, people notice their guard goes up they scroll
when it sounds like a real person
in a real space
that’s when it works
i broke down the full workflow step by step:
- create AI voiceovers that sound human
- replace fake-sounding voices (not elevenlabs)
- make audio feel native to the environment
RT + reply ''Sound'' and i’ll send the full guide
(must be following)


