01
Product shots that move
Start from a clean packshot and get a turntable, a pour, or a hand entering frame. Motion that would otherwise need a table-top studio day.
Image to Video, Free:Animate Any Photo in 1080P
Upload one still and Seedance gives it motion. The subject moves, the camera moves, the light keeps behaving, and the download carries no watermark. New accounts start with free credits.
Real output
One still, one moving scene
The image below was the only visual input. The prompt described what should happen inside it, and everything else is the model holding on to the original drawing.
Prompt
Hand-drawn comic style, three people sitting together eating fried chicken, cheerful and lively atmosphere.
The uploaded still, the only visual input.
Input image
Result
Seedance output, watermark-free.
Use cases
What a single photo turns into
Image to video is the mode for material that already exists: a product shot, an illustration, a poster, a face.
02
Illustration and character art in motion
A drawing keeps its line and palette while the scene animates, which is why comic and anime frames survive the trip better than most people expect.
03
Portraits and title beats
Give a still portrait a breath, a turn, or a look toward camera. Enough life for an intro frame or a title card.
04
Archive and print revival
Posters, scanned photos, and old campaign art become moving assets without a reshoot or a fresh rights conversation about new footage.
How it works
How image to video works here
Attaching a picture is the whole switch. The composer sees the image and routes the job, so you never pick a mode.
01
Attach the still
Drop one image into the composer and it becomes the first frame. Add a second and it becomes the last frame, so the clip travels between them.
02
Say what should move
Describe the action and the camera rather than the picture. The model can already see the frame; it needs to know what happens next.
03
Generate and keep it
Output lands in the asset library at full quality with no watermark, and can be attached again as a reference for the next pass.
Jobs people arrive with
Five image-to-video jobs, answered directly
These are the jobs people arrive with a picture for, worded the way they ask for them. Each answer says whether it works here, on which model and with what limit, then gives the steps and a prompt you can paste with a still of your own.
01
Turn a product photo into a short video ad
Yes. Attach the packshot as the first frame, ask for one product move and one camera move, and say that everything else stays as photographed. The still fixes the label, the colours and the shape, so the prompt only has to supply motion. Seedance 2.0 Mini runs it on free credits at up to 720p, enough to test the idea; Seedance 2.5 adds 1080p, sound and up to 30 seconds on a paid plan or a credit pack. The content check can refuse inputs that show well-known logos or brands; if it refuses your packshot, try a crop where the logo is smaller or turned away. A refused job is not charged.
How to do it
- Use a clean packshot: one product, a plain or simple background, sharp, at least 300 pixels on the short side.
- Attach it; it becomes the first frame and fixes how the product looks.
- Write one product action and one camera move, and say the background and lighting stay.
- Generate a 5-second version first, then rerun the one you like longer or at 1080p.
Prompt
The bottle stays centred while a slow quarter turn reveals its side. Water droplets slide down the glass, a soft band of light sweeps across the label, and the camera pushes in slightly. Background and lighting stay exactly as in the photo.
Settings
- Model: Seedance 2.0 Mini on free credits · Seedance 2.5 for 1080p and sound
- Inputs: the packshot as the first frame
- Duration: 5 seconds · Aspect ratio: adaptive, matching the photo
Prompt recipe: no clip attached, paste it and run it.
02
Animate a picture and keep the same face, outfit or product
The first frame already carries identity, palette and composition, so a short clip with a modest camera move usually holds them; drift starts with long clips and big motion. To hold a face, an outfit or a product through a whole series, switch from a single first frame to a reference group: on the Seedance models three or more stills become references rather than a fixed opening frame, up to 30 on Seedance 2.5 and nine on Seedance 2.0, Fast and Mini, and the prompt names them by number. Human faces are supported: of 6,214 image jobs run here between September 15 and October 14, 2026, only 19 were refused for showing a real person.
How to do it
- Collect two or three clean stills: front, three-quarter and full body, or the product from two angles.
- Attach them together so they act as references, and name each by number in the prompt.
- Describe the action and the camera only; do not describe the face or the product again.
- Reuse exactly the same stills for every clip in the series.
Prompt
The man in image 1, wearing the jacket from image 2, walks out of a train station into bright morning light, adjusts his collar and looks down the street. Medium tracking shot, then a slow push-in on his face. Keep his face and the jacket exactly as in the images.
Settings
- Model: Seedance 2.5 (up to 30 references) or Seedance 2.0 (up to 9)
- Inputs: two or three reference stills of the subject
- Duration: 5 to 10 seconds · keep the camera move modest for the tightest hold
Prompt recipe: no clip attached, paste it and run it.
03
Control where an image-to-video clip ends with a first and last frame
Attach two images and the job becomes a first-and-last-frame clip: it opens on the first picture, lands on the second, and the prompt describes what happens in between. Seedance 2.5, 2.0, Fast, Mini and 1.5 Pro, Kling 3.0 and O3, Veo 3.1, MiniMax H3 and Wan 3.0 take the pair; SeeDance 1 Pro Fast, Sora 2 and Sora 2 Pro take a first frame only. It is the way to make transitions, day-to-night changes and before-and-after reveals land exactly, and Seedance 2.0 Mini does it on free credits.
How to do it
- Prepare both frames at the same aspect ratio; edit the second from the first so the scene still matches.
- Attach the opening image first and the closing image second.
- Describe the change between them and end the prompt with “end on the final frame”.
- Use 5 to 10 seconds; very short clips make the change look rushed.
Prompt
The empty white living room slowly fills with furniture: a sofa slides in from the left, plants grow in the corners, a rug unrolls across the floor and the lamps switch on. Static camera. End on the final frame.
Settings
- Model: Seedance 2.0 Mini on free credits · Seedance 2.5 for 1080p
- Inputs: two stills, first and last frame
- Duration: 8 seconds · Aspect ratio: adaptive
Prompt recipe: no clip attached, paste it and run it.
04
Bring an old photo or a portrait to life
Yes, and human faces are supported. A scanned print or a painted portrait animates best when the prompt asks for small, natural motion: a blink, a breath, a slight turn of the head, hair moving in a breeze. Big gestures give the face room to drift away from the original. Seedance 2.0 Mini runs it on free credits at up to 720p. Refusals for showing a real person were rare here, 19 of 6,214 image jobs between September 15 and October 14, 2026; if a photo is refused, Kling 3.0 in the same picker is a second option, and a refused job is not charged.
How to do it
- Scan or photograph the print flat and sharp, and crop so the face is large in the frame.
- Attach it as the first frame.
- Ask for small motion only: a blink, a smile forming, a slight turn toward the camera.
- Generate 5 seconds; a longer clip gives the face more time to drift.
Prompt
Keep the photo exactly as it is, including its grain and colours. The woman slowly blinks, a gentle smile forms, and she turns her head slightly toward the camera. A light breeze moves a few strands of hair. Static camera.
Settings
- Model: Seedance 2.0 Mini on free credits · Seedance 2.5 for 1080p · Kling 3.0 as a second option
- Inputs: the photo as the first frame
- Duration: 5 seconds · Aspect ratio: adaptive, matching the photo · Audio: off
Prompt recipe: no clip attached, paste it and run it.
05
Add sound to an animated photo
Leave the audio switch on and describe the sound in the prompt. Seedance 2.5, 2.0, 2.0 Fast, 2.0 Mini and 1.5 Pro, Kling 3.0, Kling O3 and Wan 3.0 generate audio together with the motion, so a beach photo gets surf and wind and a café photo gets cups and chatter. Spoken lines need quotes and a named speaker. On the Seedance models sound does not change the credit cost; Veo 3.1, Sora 2 and MiniMax H3 have no audio option here.
How to do it
- Attach the still and keep audio on.
- Write the motion first, then two or three sounds that belong in the scene.
- For speech, quote the line and say who speaks it; keep it under ten words for a 5-second clip.
- Switch audio off and rerun if you need a silent version for your own soundtrack.
Prompt
Waves roll onto the beach in the photo and pull back over the pebbles, the palm leaves sway, and a gull crosses the sky. Sound of surf on pebbles, wind in the palms, one distant gull call. Slow push-in.
Settings
- Model: Seedance 2.0 Mini on free credits · Seedance 2.5 for 1080p
- Inputs: one still as the first frame
- Duration: 6 seconds · Audio: on
Prompt recipe: no clip attached, paste it and run it.
What you can generate
What image to video can generate here
Every row is read from the model registry the composer enforces. The inputs column says how a model takes pictures: a single first frame, a first-and-last-frame pair, or a reference group that guides identity without fixing the opening frame. The other columns are the same limits the workspace applies when you press generate.
| Model | Max resolution | Clip length | Audio | Inputs | On free credits | Starting cost |
|---|---|---|---|---|---|---|
| Seedance 2.5 | 1080p | 4–30s | On/off toggle | First frame · First + last frame · Up to 30 reference images | Paid plan or pack | From 12 credits (4s · 480p) |
| Seedance 2.0 | 1080p | 4–15s | On/off toggle | First frame · First + last frame · Up to 9 reference images | Paid plan or pack | From 9 credits (4s · 480p) |
| Seedance 2.0 Fast | 1080p | 4–15s | On/off toggle | First frame · First + last frame · Up to 9 reference images | Paid plan or pack | From 7 credits (4s · 480p) |
| Seedance 2.0 Mini | 720p | 4–15s | On/off toggle | First frame · First + last frame · Up to 9 reference images | Yes | From 5 credits (4s · 480p) |
| Seedance 1.5 Pro | 1080p | 4–12s | On/off toggle | First frame · First + last frame | Yes | From 2 credits (4s · 480p) |
| SeeDance 1 Pro Fast | 1080p | 2–12s | No audio option | First frame | Yes | From 1 credit (2s · 480p) |
| Kling O3 | 1080p | 4–15s | On/off toggle | First frame · First + last frame | Paid plan or pack | From 10 credits (4s · 720p) |
| Kling 3.0 | 1080p | 4–15s | On/off toggle | First frame · First + last frame | Paid plan or pack | From 10 credits (4s · 720p) |
| Veo 3.1 Fast | 4K | 4–8s | No audio option | First frame · First + last frame | Paid plan or pack | From 12 credits (4s · 720p) |
| Veo 3.1 Pro | 1080p | 4–8s | No audio option | First frame · First + last frame | Paid plan or pack | From 32 credits (4s · 720p) |
| Sora 2 | Model default | 4–12s | No audio option | First frame | Paid plan or pack | From 8 credits (4s) |
| Sora 2 Pro | 1080p | 4–12s | No audio option | First frame | Paid plan or pack | From 24 credits (4s · 720p) |
| MiniMax H3 | 2K | 5–15s | No audio option | First frame · First + last frame · Up to 9 reference images | Paid plan or pack | From 23 credits (5s · 2K) |
| Wan 3.0 | 1080p | 2–30s | On/off toggle | First frame · First + last frame · Up to 10 reference images | Paid plan or pack | From 2 credits (2s · 480p) |
- Free credits cover the three entry models, and all three take a first-frame image: Seedance 2.0 Mini, Seedance 1.5 Pro and SeeDance 1 Pro Fast. Any paid plan unlocks every row, and one-time credit packs unlock models by tier.
- Kling 3.0 has no image input of its own. Attaching a still quietly routes the job to its image-to-video sibling and bills it at that sibling’s rate, which is why it still appears in this table.
- A first frame, a first-and-last-frame pair or a reference group of stills does not change the Seedance rate; the job is priced by duration and resolution exactly like a text prompt, and the composer shows the figure before you commit.
- Starting cost is the catalogue rate for the shortest, lowest-resolution clip the model offers, the figure the pricing page’s comparison table quotes. Paid plans charge below catalogue, and longer or sharper clips cost more.
Resolution cap by plan
The plan sets the ceiling and the model sets what is possible under it. A 4K-capable model still renders at 720p on free credits, and Seedance tops out at 1080p on every plan.
| Plan | Seedance models | Other models |
|---|---|---|
| Free credits | 720p | 720p |
| Lite | 720p | 1080p |
| Pro | 1080p | 2K |
| Premium | 1080p | 4K |
Last updated
Cost in dollars
What an image-to-video clip costs in dollars
A first frame or a reference group does not change the Seedance rate, so these jobs cost the same as a prompt of the same length and resolution. The credits come from the calculation the generator bills with, plan discounts included, and the prices from the pricing page.
| Plan and price | Seedance 2.0 Mini5 s · 720p | Seedance 2.0 Fast5 s · 720p | Seedance 2.55 s · 720p | Seedance 2.510 s · 1080p |
|---|---|---|---|---|
| Lite, monthly$14.90/mo · 100 credits/mo | $1.7912 credits | $2.5317 credits | $4.1728 credits | Not on this plan |
| Lite, billed yearly$13/mo · 100 credits/mo | $1.5612 credits | $2.2117 credits | $3.6428 credits | Not on this plan |
| Pro, monthly$29/mo · 200 credits/mo | $1.319 credits | $1.8813 credits | $3.1922 credits | $15.37106 credits |
| Pro, billed yearly$14.50/mo · 200 credits/mo | $0.659 credits | $0.9413 credits | $1.6022 credits | $7.69106 credits |
| Premium, monthly$149/mo · 1,300 credits/mo | $1.039 credits | $1.4913 credits | $2.5222 credits | $12.03105 credits |
| Premium, billed yearly$99/mo · 1,300 credits/mo | $0.699 credits | $0.9913 credits | $1.6822 credits | $8.00105 credits |
| Starter credit pack$29.90 one-time · 200 credits | $1.9413 credits | $2.8419 credits | $4.6331 credits | $22.57151 credits |
- Each dollar figure is the plan’s price per credit times the credits that job costs on that plan, for someone subscribing today. Five seconds is the length people pick most often for image to video here, which is why three of the four jobs use it.
- Sign-up credits are not in the table: they cover a first finished clip on the entry models, Seedance 2.0 Mini, 1.5 Pro and SeeDance 1 Pro Fast, all of which take a first frame, at up to 720p and with no card on file.
- Lite caps Seedance at 720p, so the 1080p column is empty on Lite. Sound does not change the Seedance price, and a failed or refused generation is not charged.
- The composer shows the exact credit cost of the job you have set up before you press generate; if it ever disagrees with this table, the composer is right.
Measured here
How long image to video takes here, measured
Timed on this site from the moment a job is submitted to the moment the finished clip is ready, across more than 4,700 image-to-video jobs that finished between September 15 and October 14, 2026. These are real waits, not progress-bar estimates, and every row has at least 30 finished jobs behind it.
| Model | Setting picked most often | Median wait | 3 in 4 done within |
|---|---|---|---|
| Seedance 2.0 Mini | 4 s · 480p | 2 min 41 s | 3 min 8 s |
| Seedance 2.5 | 5 s · 720p | 4 min 43 s | 7 min 12 s |
| Wan 3.0 | 5 s · 720p | 3 min 14 s | 5 min 30 s |
| Seedance 2.0 | 5 s · 720p | 3 min 43 s | 5 min 17 s |
| Seedance 1.5 Pro | 5 s · 480p | 1 min 32 s | 1 min 48 s |
| Kling 3.0 | 5 s · 720p | 1 min 51 s | 2 min 32 s |
| Seedance 2.0 Fast | 5 s · 720p | 2 min 49 s | 3 min 20 s |
| SeeDance 1 Pro Fast | 5 s · 480p | 52 s | 1 min 4 s |
- A picture adds little to the wait: on Seedance 2.0 Mini the median image-to-video job took 2 min 41 s, against 2 min 22 s for a text prompt. Longer clips and 1080p take noticeably longer than the setting shown in each row.
- The wait covers everything between pressing generate and the clip appearing in your library: the upload of your image to the model provider, the provider’s queue, rendering and delivery. Unlimited-mode jobs, which can wait in a queue by design, and draft previews are left out.
- Free credits and Lite run one job at a time, Pro runs four in parallel and Premium eight, so on a paid plan several of these waits overlap.
Worked examples
Three image-to-video jobs, end to end
The first example is the real pair from the demo above with its unedited prompt. The other two are prompt recipes written against the same models and limits, one on free credits and one on Seedance 2.5, so you can paste them in with a still of your own.
01
Comic panel to moving scene
Before
One hand-drawn, comic-style still of three people sharing fried chicken. It was the only visual input, and the prompt restated the style rather than describing new motion in detail.
After
The same drawing in motion: the characters eat and react while the line work and the palette hold, and the composition stays where the artist put it. This is the clip shown in the demo section above.
Prompt
Hand-drawn comic style, three people sitting together eating fried chicken, cheerful and lively atmosphere.
Settings
- Model: Seedance 2.0
- Inputs: one still as the first frame
- Entry: image to video, inferred from the single attachment
02
Packshot turntable on free credits
Before
A clean e-commerce packshot: a ceramic pour-over coffee dripper on a white sweep, front-lit, perfectly still, the kind of image every product listing already has.
After
A five-second clip in which the dripper turns a slow quarter rotation on the spot while a hand enters from the right and pours water, steam rising; the background, lighting and framing stay exactly as photographed.
Prompt
The ceramic dripper rotates slowly a quarter turn on the spot. A hand enters from the right and pours hot water in a thin stream, steam rising. Background, lighting and framing stay exactly as in the image.
Settings
- Model: Seedance 2.0 Mini, available on free credits
- Inputs: the packshot as the first frame
- Duration: 5 seconds · Aspect ratio: adaptive, matching the still
- Resolution: 720p · Audio: off
Prompt recipe: no clip attached, paste it and run it.
03
Day-to-night poster with a first and last frame
Before
Two stills: a vintage travel poster of a seaside town, and a version of the same poster with the sky darkened and the windows lit, prepared in an image editor.
After
A ten-second clip that travels from the day version to the night version: clouds drift, the sea moves, lights come on one by one, and the clip lands exactly on the second frame because it was pinned as the last one.
Prompt
Slow transition from day to evening over the seaside town. Clouds drift right, waves roll gently, windows light up one by one, the sky deepens to indigo. End precisely on the final frame. Subtle camera push-in, illustrated poster texture preserved.
Settings
- Model: Seedance 2.5, paid plan or credit pack required
- Inputs: two stills as first and last frame
- Duration: 10 seconds · Aspect ratio: adaptive
- Resolution: 1080p on Pro and Premium, 720p on Lite · Audio: on
Prompt recipe: no clip attached, paste it and run it.
Limits
Where image to video falls short, and why a job gets refused
Knowing the failure modes saves credits and retries. The refusal figures are counted from the image jobs run on this site between September 15 and October 14, 2026.
01
Logos, brands and known characters in the picture
This is the refusal people hit most with a picture. About one image job in seven on Seedance 2.0 Mini and Seedance 2.5 was refused by the content check, and the reason it names most often is copyrighted or trademarked material: a prominent third-party logo, a famous product design, or a character from a film, game or cartoon. Crop the logo out or use your own artwork; describing a style is fine, uploading the character usually is not. A refused job is not charged.
02
Flagged images, and the rare real-person refusal
The next most common refusal is an input image the check flags as sensitive. Faces themselves are supported: only 19 of 6,214 image jobs here were refused for showing a real person. If a photo keeps being refused on Seedance, Kling 3.0 in the same picker is a second option; it refused none of its 63 image jobs in the same period.
03
Small or soft source images
The upload must be at least 300 pixels on the short side, and a small or heavily compressed still stays soft even when the output is 1080p. Start from the sharpest version you have; the clip cannot add detail the photo never had.
04
Big motion breaks the likeness
The further the subject moves from the first frame, the more fine detail drifts: faces, hands and small print on a label. Keep the camera move modest, describe one action, or pin the ending with a last frame so the clip has to land on a known picture.
05
Labels and text once they move
Text that is crisp in the still can warp as soon as the product turns or the camera moves. Keep the label facing the camera, or add the lettering on top in your editor afterwards.
06
Collages and busy frames
Crops that cut the subject in half, collages and frames crowded with small objects give the model less to hold on to, and the motion goes where you did not ask for it. One clear subject on a simple background animates best.
Why Seedance refuses prompts, counted on 11,168 jobs, with rewrites
Questions
Image to video questions
- Which images work best, and what are the upload rules?
- A sharp, clean still with an obvious subject. Crops that cut the subject in half, heavy compression, and busy collages give the model less to hold on to. On Seedance 2.0, 2.0 Fast and 2.0 Mini the rules are JPEG, PNG or WebP up to 30 MB, at least 300 pixels on the short side and no more than 6,000 on the long side, with an aspect ratio between 2:5 and 5:2; Seedance 2.5 takes the same formats at the same 30 MB and 300-pixel floor. The composer rejects anything outside those limits before a credit is spent. Only upload images you have the right to use.
- Is it free, is there a watermark, and do I need a card?
- New accounts start with free credits and no card, and the grant is sized for a first finished clip: two generation attempts and one completed video before a plan or a credit pack is needed. Free credits run the entry models, Seedance 2.0 Mini, Seedance 1.5 Pro and SeeDance 1 Pro Fast, all of which take a first-frame image, at up to 720p. Nothing rendered here is watermarked at any tier, so the download is the clean file.
- Which models animate an image, and can I pin where the clip ends?
- Every model in the table above accepts a first frame: the six Seedance models, Kling 3.0 and Kling O3, Veo 3.1 Fast and Pro, Sora 2 and Sora 2 Pro, MiniMax H3 and Wan 3.0. To fix the landing point, attach two images and they become the first and last frame; Seedance 2.5, 2.0, Fast, Mini and 1.5 Pro, Kling 3.0 and O3, Veo 3.1, MiniMax H3 and Wan 3.0 all take the pair, while SeeDance 1 Pro Fast, Sora 2 and Sora 2 Pro accept a first frame only. On the Seedance models, three or more images become a reference group instead.
- What resolution and length can an animated photo reach?
- Length follows the model: 4 to 30 seconds on Seedance 2.5, up to 15 on Seedance 2.0, Fast and Mini, Kling 3.0 and O3 and MiniMax H3, up to 12 on Sora 2 and Seedance 1.5 Pro, up to 8 on Veo 3.1, and 2 to 30 on Wan 3.0. Resolution is capped by the plan before the model: 720p on free credits, then for non-Seedance models 1080p on Lite, 2K on Pro and 4K on Premium where the model offers it. Seedance itself stops at 1080p, which Pro and Premium unlock. The source image does not have to match the output size, but 1080p output from a small upload will look soft, so start from the sharpest still you have.
- How many credits does animating a photo cost?
- The starting-cost column shows each model’s catalogue rate at its shortest, lowest-resolution setting, before plan discounts; the composer prices the exact job before you press generate. A first frame or a reference group of stills does not change the Seedance rate, and on Kling 3.0 the job is billed at the image-to-video sibling’s rate. Longer and sharper clips cost more, every paid plan charges below catalogue, and failed generations are not charged. For dollars, the cost table above converts four common jobs at each plan’s price per credit.
- Is seeddance.io the official ByteDance site for Seedance?
- No. SeedVideo AI at seeddance.io is an independent platform, not a ByteDance property. It runs the official Seedance model API alongside Kling, Veo, Sora, MiniMax and Wan, so the Seedance output you get here is produced by the model ByteDance publishes, but the workspace, the credits and the plans are ours. Model names belong to their respective owners, and nothing on this site is endorsed by them.
- Can I use image-to-video output commercially?
- Commercial use is included on Pro, Premium and the annual Lite plan, and not on free credits or the monthly Lite plan. The file is identical across tiers; the licence differs. Because this entry starts from your own picture, the rights question is mostly about the input: product photos, artwork and portraits you upload must be yours to use, and a licence from us does not cover a photograph you do not own.
- How long does an image-to-video generation take?
- A few minutes. Measured on this site over the 30 days to October 14, 2026, the median image-to-video job finished in 52 s on SeeDance 1 Pro Fast, 1 min 32 s on Seedance 1.5 Pro, 1 min 51 s on Kling 3.0, 2 min 41 s on Seedance 2.0 Mini, 2 min 49 s on Seedance 2.0 Fast, 3 min 14 s on Wan 3.0, 3 min 43 s on Seedance 2.0 and 4 min 43 s on Seedance 2.5, each at the setting people pick most often; the measured table above has the full figures. Longer clips, 1080p and busy periods stretch them, and the workspace shows an “about N min left” estimate. Free credits and Lite run one job at a time; Pro runs four in parallel and Premium eight.
- Why was my image-to-video job refused?
- Almost always because of the model’s content check rather than a fault. Between September 15 and October 14, 2026, about one image job in seven on Seedance 2.0 Mini and Seedance 2.5 was refused, and the reason named most often is copyrighted material, such as a prominent logo, a brand or a known character, followed by images the check flags as sensitive. Refusals for showing a real person were rare: 19 of 6,214 image jobs. Crop the logo out, use your own artwork, or try the same image on Kling 3.0, which refused none of its 63 image jobs in the same period. A refused job is not charged.
- How do I keep the product or character consistent from the still?
- The still already does most of the work: identity, palette and composition carry over from the first frame, while fine detail can drift as motion increases. To hold them through a longer or busier clip, keep the camera move modest, describe the action rather than re-describing the picture, and use the first-plus-last-frame pair so the ending is pinned too. For a series, attach the same product still as a reference on every generation, up to 30 references on Seedance 2.5 and nine on Seedance 2.0, and feed finished clips back in as references for the next pass.
- Image to video or video to video for a product?
- If all you have is a photo of the product, use image to video: the still becomes the first frame and the prompt supplies the motion. If you already have footage of a similar product, video to video on Seedance 2.5 keeps the set, the lighting and the camera and swaps the object for the one in your still, which is usually faster than staging the shot again from a single frame. When both are available, animate the packshot for hero motion and edit the footage for catalogue variants.
Choosing an entry
Image, text or video: which entry fits
The three scene pages share one composer. What you attach decides the entry, and the model you pick decides the ceiling.
A still is the right entry when the look is already decided and the job is to give it motion. When nothing exists yet and you want to explore quickly, start from text to video and write the shot instead. When you have footage that is almost right, use video to video and describe only the change, so the camera, set and timing you already like stay put.
The model choice is separate from the entry. Seedance 2.5 is the pick when the still has to survive a long clip or travel through several shots, because it holds up to 30 reference images and runs to 30 seconds. Seedance 2.0 Mini is the pick for a quick turntable on free credits, and the first-and-last-frame pair works across most of the table when the ending matters as much as the start.
When the clip has to land on a specific picture, attach the closing image as well: first and last frame to video lists the 11 models that take the pair, what each charges per second, and a worked example with both frames and the full instruction.
Keep going
Where to go next
A still is one of three ways in. Pick the entry that matches what you already have.
Free credits
Upload a still, get a scene
Drop in one photo, say what should move, and download the take without a watermark. Free credits on a new account.
Animate an image