Twitter 社区 · Seedance 提示词案例

This is the direction AI video has needed — real footage as the control layer, not prompt roulette. Seedance takes drone footage of a real hometown and turns it into impossible VFX shots. Shoot a plate → direct the AI. Practical, controllable, production-ready.

Real footage is becoming the control layer for AI video. This Seedance clip starts from drone footage of a real hometown, then turns the same location into increasingly impossible tsunami VFX. That feels more practical than another prompt-only demo: shoot a plate, keep the

Good Morning and Happy Payday Everyone. Seedance 2.0 prompt 👇 15 seconds, single continuous handheld selfie vlog POV, energetic pacing, no hard cuts. Global setup: High energy viral content-creator video. Wide angle vlog lens, natural handheld shake, bright vibrant lighting

created by seedance 2.0 mini in xiaoyunque doubao. great prompt.

This is an older video, but I don't think I’ve shared it here yet. It was generated back in late February during the Seedance 2.0 early access. At the time, I was already using an "auteur script" structure, but with different block names: GLOBAL > LOOKBOOK/REFERENCE > SCREENPLAY.

Fable 5 vs Sonnet 5 on the same video model (Seedance 2.0). The real difference? How each interprets and structures prompts in an agentic pipeline. That single choice determines your generation success rate per run. Which one has been more consistent for you? 🤔

Someone filmed a full day of student life in Jakarta. Except no one filmed anything. This is Seedance 2.0. One prompt. 8 shots. Zero cameras. If this doesn't scare you, you're not paying attention. Comment "PROMPT" and I'll send you the exact prompt.

Reference Video → 4K Video · Seedance 4K - Feed in your video as a reference video. - Seedance adds photorealistic VFX in true 4K matching your composition and camera movement exactly.

3D → 4K Video · Blender + Seedance 4k - Convert 3D, CG, or game renders into photorealistic, film-quality footage. - Uses your render as a motion and framing reference to preserve exact composition and camera movement while upgrading to high-fidelity realism.

Cinematic battle scene between Ren (Mind Dominator) and Nox (Visual Illusionist). Mind control energy vs blinding light distortions, epic showdown, highly detailed, 8k. Image: Midjourney & GPT image 2 Video: seedance 2.0 @dreamina_ai Have a great Friday! #dreaminacpp

A grotesque monster inspired by John Carpenter's the Thing. 80s style analog filmed. Cool color grade. Desaturated. https://t.co/HTCLnVy0KY

Everyone's blocking their Seedance shots in Blender. We're building a tool more streamlined: PINOC Studio, a control layer for AI video. Block the scene, set every camera. Next step: we're adding character motions! It's not live yet but soon. Comment & retweet, we'll DM you

Seedance 2.0 mini is now supported in ZooClaw: fast video generation, built for high-volume content, with ~50% lower token cost. You watch the game. ZooClaw ships the videos.

Visual Preset #01 Ink-Brush Cinematic 3D: A high-end cinematic 3D style where expressive ink-brush effects become the primary visual language for fast-paced anime action. Lately I've noticed that I've been experimenting with different visual presets across my videos and I'd like

@YouArtStudio It's hard to create a seedance prompt for these

15 seconds, 16:9 cinematic, single continuous handheld shot following the character, minimal cuts (one hard cut at the hook moment for shock value), unhurried until the final beat. Global setup: https://t.co/urFdMPMLH5

Kling AI, the AI video generation spinoff from Kuaishou Technology, has raised an initial $2 billion in venture capital funding to expand its AI video operations, with the round potentially growing to about $3 billion as additional investors join. Valued at $15 billion before the

Midjourney v8.2 and Seedance 2.0 at 4K. Stunning images combine well with amazing video tools.

Just watched this breakdown and it’s one of the cleanest examples of how AI is quietly creating new money printers in content. The video shows exactly how to make those viral “You vs Future You” cinematic videos using Seedance 2.0 (ByteDance’s new multimodal model). Two

She's not real. The performance kind of is. 12-sec screen test: one prompt, three cuts, lip-synced line, native room tone. Seedance 2.0 via @magnific, 1080p, no edit. Trick: write cuts as prose, keep dialogue under 12 words. Cast her — yes or no? #AIVideo #Seedance #magnific

15 seconds, 16:9 cinematic, single continuous slow tracking shot down a hallway, deliberately unhurried, minimal camera movement, no cuts. Global setup: Prestige horror film aesthetic, inspired by https://t.co/huGe9ViY07

MOST CREATORS PROMPT SEEDANCE 2.0 WRONG The fix takes 12 minutes and one JSON block. Here's the exact pipeline: Step 1. Generate a base frame in ChatGPT Image. Keep it specific: subject, lighting, camera distance, mood. Step 2. Drop that frame into Seedance 2.0 as the

You all really loved Sniper Girl from the last video, and I do give the people what they want! Today I'm diving into the latest Seedance 2.5 Released Videos, AND: A Beta Bombshell! Link Below!

Experimental motion design Seedance 2.0 Fast on @magnific. Prompt below: