AI Agent 就绪，开放 API — 支持 Claude Code、Codex、OpenClaw 等
现在订阅，Seedance 2.0 额外享 8 折优惠，所有付费套餐通用
额外赠送 Kling V3 2026 年 7 折优惠
锁定即将上线的 HappyHorse 模型 8 折优惠
Built for AI Agents, end to end.
共漫Convibe ships a full REST API plus the skillsvideo agent-native CLI. Every provider, every model — callable from HTTP, any script, or dropped straight into an agent runtime.
$skillsvideo generate "a corgi astronaut, neon city"
submitting to seedance-2.0 · 24s
✓ ready https://skills.video/v/k3p9aq
Ships with first-class support for the agents you already run.
Claude Code
Anthropic's terminal-native coding agent.
OpenAI Codex
OpenAI's autonomous code agent for your shell.
OpenClaw
The AI that actually does things.
Hermes Agent
An agent that grows with you.
Nexu
Open-source OpenClaw desktop app for Feishu.
Spacebot
An agentic system with specialized processes for thinking, working, and remembering.
Seedance 2.0 prompt guide
Build the prompt from subject, action, scene, camera, and audio
Seedance 2.0 and Seedance 2.0 Fast follow natural-language instructions, so write the prompt as a compact shot brief instead of a keyword pile. Start with who or what appears, what they do, then add environment, visual style, camera movement, and sound when those details matter.
Required core: subject plus action, so the model knows who is doing what.
Optional control: scene, lighting, style, camera direction, pacing, and sound.
Use one main action per clip before adding secondary details.
Prompt example
Close product shot of a black wireless speaker on wet concrete, water droplets vibrate with the bass, slow dolly-in, blue hour city lights, realistic reflections, low cinematic synth.
Name each reference image, video, or audio source clearly
The official guide emphasizes explicit reference control. Upload assets in the right order, then point to them by name in the prompt, such as image 1 for composition, image 2 for a character, or video 1 for motion and camera language.
Use image references for character identity, product details, logos, scene layouts, and storyboard panels.
Use video references when the action, camera path, or visual effect needs to stay consistent.
Say exactly which part of each asset should be copied and which part can change.
seedance 2.0 image to videoseedance image to video prompt guideseedance 2.0 reference video
Control on-screen text, subtitles, and speech bubbles
Seedance 2.0 can generate text in text-to-video, image-to-video, reference-to-video, and video-to-video workflows. For stronger results, specify the wording, timing, screen position, display style, and whether subtitles should sync to the voice.
For slogans, include the exact text, where it appears, when it appears, and the visual treatment.
For subtitles, ask for bottom captions that follow the dialogue or narration rhythm.
Prefer common characters and simple punctuation when text legibility matters.
seedance 2.0 text to videoseedance 2.0 subtitlesseedance 2.0 ai video generator
Use video reference and edit prompts for advanced control
For video-to-video work, Seedance 2.0 supports prompts that reference existing motion, camera moves, effects, and edit operations. Write the instruction around the source video first, then define what should be added, removed, replaced, extended, or connected.
For action transfer, describe the action in video 1 and the new subject or scene.
For camera transfer, describe the camera move from the reference video and the target scene.
For editing, specify the time position, spatial position, target element, and what should stay unchanged.
seedance 2seedance 2.0how to use seedance 2.0seedance 2.0 video editing
