Twitter 社区 · Seedance 提示词案例

I’ve been trying to make a ballet music video for my song “You” for a while now, but every time I tried, something just didn’t come out quite right. Then Seedance 2.5 dropped… and somehow on DAY ONE it finally worked. My song "You" is on Spotify 💙

First Seedance 2.5 generation on @YouArtStudio! Some notes: 🎵 The music. There's a music bed at the top of this clip that we didn't ask for. "No music" is explicitly in the prompt. Other people have flagged the same thing on 2.5, so it seems to be a known quirk rather than a us

Seedance 2.5 is now available in Novoads!! Dm me if you want to know how I made the video 😊

K-pop boyband. Seedance 2.5 made one anyway. 7 warriors. one drakkar. flawless choreography on drifting ice. not a single one of them is smiling. 30 seconds, one prompt, one generation. sound is native.

I think we’re finally at the point where the emotional depth of AI video is actually watchable, at least for short-form clips. The technical progress here is kind of ridiculous too, Seedance 2.0 only launched 176 days ago, and now 2.5 is already pushing longer single shot

Seedance 2.5 is live on 4 to 30 second shots with first and last frame control, plus image, video and audio references when continuity matters.Try it:

SYLVANORA — "The Veiled Pas De Deux" | Opening Titles. 15-second montage, 30 beats @ 0.5s each, fluid acrobatic dancer's pacing. Art Direction & Vibe: Premium key art. Unified style locked to reference: highly detailed dark fantasy digital concept art, fine https://t.co/2dWXfEq9Z3

New toilet world video made with Seedance 2.5. Disturbing. #toiletworld

TOPVIEW AI（ @TopviewAIJP ）で昨日リリースされた Seedance 2.5の比較動画を、まずは1本目として作ってみました。今回は前モデルである Seedance 2.0 との T2V（Text to Video）比較になっています。

We’re incredibly impressed by Seedance 2.5’s ability to turn a single image and a detailed prompt into a coherent 30-second edit. The fast cuts, additional scenes, transition styles, and expressions were exactly what we had envisioned.

30 -seconds Seedance 2.5 video prompt share👇 Created with Seedance 2.5 @wavespeed_ai @HiWscreator #Transformers #Prompt #Grok

三国志10週くらい回ってやっぱり関羽がかっこいい！！ ナレーションあってるかどうかわからんｗｗ 【PR / SeaArt Creator Mission】 30s AI video generation is here, created with Seedance 2.5 on SeaArt. @SeaArt_Ai @SeaArtCoLab #SeaArtSeedance25

Fight scene test in Seedance 2.5 in @TopviewAIhq Viper the Assassin vs Raina the Monk Sharing references and prompt in reply! ⬇️⬇️⬇️

Seedance 2.5 is LIVE on Higgsfield UNLIMITED for 33 days at zero credit cost. You'd be INSANE to waste this window: 1. Seedance 2.5 dropped on Higgsfield today 2. Unlimited generations for 33 days 3. Zero credit cost. Read that again. 4. No rationing. No credit anxiety. No

Hey there everyone! I'm the Chaos Angel Sakura! I'm not your normal Giantess Kaiju Vtuber! PSST.. Normally we don't do TWO major posts in one day but I had to share this video with you all made with MY voice actress modeling it & using Seedance 2.5! Made with @kinovi_ai !

Loved how Flux breaks glass, so I went ahead and smashed even more. 🪟 😈 Seedance 2.5 ⚔️ Flux 3, same prompt via @krea_ai

cat swimming with jellyfish The algorithm is working - I'm getting more cat things on my timeline. --------------------------------- Grok water, reflections, and jelly physics is pretty good but cat seems to be redrawn https://t.co/VsQskJShj1 https://t.co/HFHd7ZVTUM

SEEDANCE 2.5 Philippines: PINOY AI FILMMAKER | Official Music Video #Seedance25 #Seedance20 #PinoyAIFilmmaker #AIMusicVideo #OfficialMusicVideo #SunoAI #OpenArtAI #CapCutVideoStudio #CapCutAI #AIVideo #AIContentCreator #CinematicAI #AIStorytelling #GenerativeAI #FilipinoCreator

Seedance fight scenes are getting past the “pretty clip” stage. The signal is contact timing, camera rhythm, readable choreography, and bodies that do not collapse when the action gets fast. That is where AI video starts to feel like direction, not only generation.

We need some velociraptor prompt Seedance 2.0 @openart_ai The Bride Hyper-realistic cinematic comedy sequence, 12 seconds, aspect ratio 16:9. Daytime inside a luxury bridal boutique. One female velociraptor only. The setting is simple and clear: white wedding dresses, mirrors,

Made with Seedance 2.0 by accident, no reference image, no complete prompt just a "Cinematic 3D CGI Style Prompt Block" and I got this crazy result. I mistakenly pressed enter 2 times inside prompt box while making a new line in between the style blocks and got 2 videos

I am vibecoding a simple app, i called Scene Blocker, to help me block character placement and movement inside a shot, position the camera, and use a timeline to keyframe the action. Then I can export that blocking as a video, combine it with character and environment references,

TWO AI VIDEO MODELS GOT THE SAME PROMPTS. I WATCHED BOTH ON LOOP FOR AN HOUR AND I STILL CAN'T PICK A WINNER no commentary. no voiceover. just both models. same prompts. side by side i went in expecting a clear winner i was wrong MiniMax H3 - the motion is cinematic. the way

The Zephyr blueprint is out in the open million+ views in hours they made every character, prompt and asset from the series public to study and remix you can build the sequel, send the cast somewhere unexpected with Unlimited Seedance offer Part of the Higgsfield Global Film