THE CREATOR’S NOTEBOOK
Make your next video with Seedance.
Prompts, reference images, video settings and fixes for your next shot.
CHARACTER CONTINUITYKeep a Character Consistent in Seedance — Reference Image Guide
To guide the same character across Seedance videos, attach the same clear reference image to each shot, keep the outfit and visual style unchanged, and describe the new action. Review the face and silhouette in every result: a reference guides continuity but does not lock a character perfectly.
Read the guide
VIDEO TROUBLESHOOTINGSeedance Video Troubleshooting — Motion, Quality & Failed Jobs
Check the task status before spending credits on another attempt. For a completed Seedance video, identify whether the problem is subject movement, camera movement, the input mode or changing details. Use the checks below to plan one specific revision and record what changes.
Read the guide
Getting startedHow to Use Seedance 2.5: Text and Image to Video
Start with one clear shot. This guide takes you from an idea or a single image to a downloadable video in SuperSeedance.
Read the guide
PROMPT CRAFTSeedance Prompt Guide: Motion, Camera & 8 Video Examples
Write a Seedance prompt as one visible action: who moves, where they go, how the camera follows and what must stay consistent. Start from eight generated motion examples, then adapt the action and visual style to your own shot.
Read the guidePRODUCT WORKFLOWTurn a Product Photo into a Video Ad — Workflow
Turn one product image into a video shot, then use the on-site AMBER template for a five-second 9:16 silent ad with editable brand copy and CTA. Text revisions reuse the original. The historical case below preserves the separate external edit with sound, its original clip and downloadable project.
Read the guide
COST & BUDGETSeedance Video Cost — Credits & Budget Guide
A five-second Seedance 2.5 text-to-video clip at 480p, 16:9, with audio and draft mode off uses 83 credits on SuperSeedance. Longer clips and higher resolutions use more. Check the quote for your exact settings and budget for every completed attempt.
Read the guideMODEL DECISIONSHow to Choose a Seedance Model — Decision Framework
Start with your scene, output format and budget. This guide helps you choose useful video settings and compare results that matter to your project. Seedance 2.5 is currently available on SuperSeedance.
Read the guideCAMERA WORKFLOWSeedance Camera Prompts: Fixed Shots, Tracking and Orbits
To direct the camera in Seedance, describe its starting view, path and ending separately from the subject’s action. Choose a fixed view, a tracking move or a short orbit for one shot. SuperSeedance takes these directions in the prompt; it has no camera-path slider or guaranteed camera lock.
Read the guideVERTICAL VIDEOHow to Make Vertical 9:16 Videos with Seedance 2.5
For a vertical Seedance video, choose 9:16 in text or reference mode. In first-frame image mode, prepare a 9:16 source image instead: Seedance 2.5 uses Match image. Then plan the action for that narrow frame and check the complete downloaded clip before publishing.
Read the guideAUDIO WORKFLOWHow to Make Seedance Videos with Sound
To make a Seedance 2.5 video with sound in SuperSeedance, turn on Native Audio before generating and describe the sound alongside the scene. New workspaces start with audio off. For complete silence, leave it off; for dialogue or sound effects without music, keep it on and describe which sounds you want.
Read the guideSHORT FILM WORKFLOWMake a Longer Seedance Video: A Three-Shot Workflow
To make a Seedance video longer than 30 seconds in SuperSeedance, generate separate clips, download them and assemble them in a video editor. Start with a three-shot plan: one arrival, one crossing and one reveal.
Read the guideIMAGE WORKFLOWFirst Frame vs Reference Image in Seedance 2.5
In our two Seedance 2.5 requests using the same image and prompt, the first-frame output kept the whole bottle visible in the reviewed frames; the reference output cropped the cap. Compare the originals below. The seeds and delivered ratios differ, so this pair does not establish that one mode is generally better.
Read the guide