Text to video and image to video on ByteDance’s Seedance 2.5 — clips up to 30 seconds with sound, right in your browser.
Native audio · Text & image to video · Up to 30s · Up to 1080p
Background: real output · Seedance 2.5 Flash · 480p · 15s · audio on
Seedance 2.5 is ByteDance’s newest AI video model, launched on July 31, 2026. It creates picture and sound in one pass and can tell a story in a single 30-second take. Seedance25Video runs it online: write a prompt or upload a photo, pick Seedance 2.5, Turbo or Flash, and download an MP4 with audio. Plans start at $14.90 a month.
Seedance 2.5 Flash is the lowest-cost tier. Switch to Turbo or the full model for 720p and 1080p.
What went in and what came out: the exact prompt (and start image) behind each clip. Copy one, or load it into the generator and make it yours.
Reference
Final result
Start image
Seedance 2.5 Flash · image to video · 480p · 5s
Prompt
The street chef laughs, wipes his hands on a towel over his shoulder and waves at the camera as steam drifts past the string lights. Crowd chatter and sizzling woks.
Reference
Final result
No upload
Seedance 2.5 Flash · 480p · 15s
Prompt
A three-shot story at dusk: a girl in a yellow raincoat runs down a rainy Tokyo side street, neon signs glowing; cut to her ducking under a ramen shop awning and shaking the water off her hood; cut to a warm close-up as she smiles at the cook and says "One bowl, please." Rain on the awning, sizzling broth, soft city noise.
Reference
Final result
No upload
Seedance 2.5 Turbo · 720p · 5s
Prompt
Vertical handheld shot at a crowded night market: a young street chef tosses noodles in a flaming wok, then looks up at the camera, grins and says "Order up!". Sizzling oil, clanging pans, crowd chatter.
Reference
Final result
No upload
Seedance 2.5 Flash · 480p · 5s
Prompt
A lighthouse keeper climbs a spiral iron staircase at night, his lantern swinging; cut to the lamp room as the great beam sweeps across a stormy sea. Thunder rumbles, waves crash against the rocks.
Real, unedited Seedance 2.5 outputs with audio on — wide, vertical and photo-to-video. On a computer, hover a clip to load its prompt into the generator.
What Seedance 2.5 does well, and what to lean on when you write prompts for it.
Seedance 2.5 generates up to 30 seconds in a single pass, twice the 15-second limit of Seedance 2.0, so a scene keeps its characters, light and sound. Write the beats in order (“cut to the lamp room”) and they play as shots. A 30-second Flash clip at 480p costs 1260 credits.
Speech, ambience, sound effects and music are generated with the motion, so the pour, the footsteps or the rain land on the right frame. Put spoken lines in quotes (the chef says “Order up!”), or switch audio off for silent footage.
Switch to Image → Video and upload a picture: Seedance 2.5 keeps its subject and composition and adds motion and sound from your prompt. Add an optional last frame to control where the shot ends. The clip keeps your image’s framing.
Six aspect ratios for text to video: 9:16, 3:4, 1:1, 4:3, 16:9 and 21:9. Generate 9:16 directly for TikTok, Reels and Shorts instead of cropping a widescreen clip, or 21:9 for a cinematic frame.
The full Seedance 2.5 model for detail up to 1080p, Turbo for HD at a lower price and Flash for the cheapest drafts and long takes. The credit price updates as you change the model, length or resolution.
Credits are reserved when a clip starts and returned automatically if the generation fails. Every finished Seedance 2.5 video is saved to your history and downloads as an MP4 with its audio track.
ByteDance’s video model for 30-second storytelling with sound — and you don’t need a ByteDance account to use it here.
Seedance is the AI video model family from ByteDance’s Seed team. Seedance 2.5 was previewed at the Volcano Engine FORCE conference on June 23, 2026 and launched on July 31, 2026. ByteDance describes it as “a next-generation audio-video joint generation model, built for 30-second storytelling with precise reference control and powerful editing capabilities.”
Compared with Seedance 2.0, version 2.5 doubles the length of a single generation from 15 to 30 seconds, follows reference footage more precisely and, in ByteDance’s words, makes motion “smoother and more consistent” with more realistic visuals. Because picture and sound are generated together, footsteps, rain, voices and music land on the right frame instead of being added afterwards.
Seedance25Video is an independent web app that runs Seedance 2.5 through an API provider. You get three tiers in one Seedance 2.5 video generator — the full model, Turbo and Flash — with text-to-video, image-to-video, clips from 5 to 30 seconds and output up to 1080p.
What changed between the two versions, from ByteDance’s announcements and the API specs.
| Seedance 2.0 | Seedance 2.5 | |
|---|---|---|
| Released | February 2026 | July 31, 2026 |
| Longest single generation | 15 seconds | 30 seconds |
| Reference inputs | Up to 9 images, 3 videos and 3 audio clips | Up to 30 images, 10 videos and 10 audio clips |
| Audio | Generated with the picture | Generated with the picture |
| Motion and editing | Film-like motion | Smoother, more consistent motion; more reliable editing |
| API price at 720p | $0.24 per second | $0.36 per second (Turbo $0.20, Flash $0.24) |
Reference inputs and video editing are part of Seedance 2.5 but not available on this site yet.
All three are versions of ByteDance’s Seedance 2.5 with native audio and up to 30-second clips. They differ in resolution and price per second.
The full model. Use it for final shots, close-ups and scenes where detail and motion matter most.
A faster, cheaper tier that outputs 720p or 1080p. 1080p on Turbo costs a fraction of the full model.
The lowest-cost tier at 480p or 720p. Ideal for trying prompts, long 30-second takes and high volume.
Seedance 2.5 is the star here, but every model below has its own generator and uses the same credit balance.
Need a still first? Generate a start image withNano Banana ProNano Banana 2.1Seedream 5.5GPT Image 2
Generate 9:16 clips for TikTok, Reels and Shorts directly — hooks, B-roll and story beats with the sound already on the track, so a Seedance 2.5 clip can go straight into your edit.
Test several versions of one idea before you shoot anything. Change the setting, the light or the line of dialogue in the prompt and compare the takes side by side.
Upload a product photo as the first frame and describe a slow turntable, a pour or an unboxing. Short loops work well on product pages and marketplace listings.
Block out a scene before the shoot: test framing, lens feel and lighting, or use a 30-second take to rough out a whole sequence with its sound.
Animate a landscape, a portrait or an illustration from a single image — only with pictures you own or have permission to use.
Create mood shots and visual loops for songs, podcasts and stories. Mention the sound you want — rain, a crowd, wind, a melody — and it is generated with the picture.
From idea to a downloadable clip in three steps — all in the generator at the top of this page.
Write what happens: the subject, the action, the camera move, the light and the sound. To animate a picture, switch to Image → Video and upload it as the first frame (and, if you like, a last frame).
Choose Seedance 2.5, Turbo or Flash, then the length (5, 10, 15 or 30 seconds), the resolution and the aspect ratio. The credit price updates as you change them.
A 5-second clip usually takes a few minutes; 1080p and 30-second videos take longer. Finished videos are saved to your history as MP4s with audio, and failed ones are refunded automatically.
Credits per clip on every Seedance 2.5 tier — the same prices the generator above charges. One credit is about one US cent.
| Model | Resolution | 5s | 10s | 15s | 30s |
|---|---|---|---|---|---|
| Seedance 2.5 | 480p | 315 | 630 | 945 | 1890 |
| 720p | 630 | 1260 | 1890 | 3780 | |
| 1080p | 1575 | 3150 | 4725 | 9450 | |
| Seedance 2.5 Turbo | 720p | 350 | 700 | 1050 | 2100 |
| 1080p | 385 | 770 | 1155 | 2310 | |
| Seedance 2.5 Flash | 480p | 210 | 420 | 630 | 1260 |
| 720p | 420 | 840 | 1260 | 2520 |
Credits are only charged for finished videos; failed generations are refunded. Plans start at $14.90 a month, credits included.
Seedance 2.5 can do more than this site offers today. Here is exactly what is available.
Seedance 2.5 accepts up to 30 reference images and 10 videos and audio clips per generation. This site does not support reference uploads yet; image to video takes a first and last frame.
Editing an existing clip and extending it are Seedance 2.5 features we have not added. Every generation here is a new clip from a prompt or an image.
Seedance 2.5 runs at 480p, 720p and 1080p here; 4K is not offered for it. If you need 4K, Seedance 2.0 and Veo 3.1 offer it in their own generators, at a higher price per second.
Don’t upload photos of people without their consent, and don’t recreate celebrities or copyrighted characters. Requests the model provider rejects fail without charge.
Seedance 2.5 is ByteDance’s AI video generation model, launched on July 31, 2026. It creates video with synchronized audio from a text prompt or an image, in clips of up to 30 seconds.
Not for video. Making Seedance 2.5 videos needs a plan, from $14.90 a month with 1,800 credits included — about 8 Seedance 2.5 Flash clips at 480p and 5 seconds. New accounts get 210 free credits to try the image models, no card required.
A 5-second clip costs 210 credits on Flash at 480p, 350 credits on Turbo at 720p, 385 credits on Turbo at 1080p and 1575 credits on the full Seedance 2.5 model at 1080p. Longer clips cost proportionally more. One credit is about one US cent.
The full Seedance 2.5 model gives the best detail and offers 480p, 720p and 1080p. Turbo outputs 720p or 1080p for much less per second. Flash is the cheapest tier at 480p or 720p. All three generate audio and support 5- to 30-second clips.
You can choose 5, 10, 15 or 30 seconds. Seedance 2.5 generates the whole clip in one pass, so a 30-second video is one continuous take, not stitched segments.
Yes. Set the generator to Image → Video and upload a picture as the first frame; you can also add a last frame. Seedance 2.5 animates it following your prompt and keeps the image’s aspect ratio.
Yes. Audio is generated together with the video: ambience, sound effects, music and speech. Put spoken lines in quotes in your prompt. You can also switch audio off for silent clips.
A 5-second clip usually takes a few minutes. 1080p and 30-second videos can take 10 minutes or more. You can leave the page — finished videos are saved in your history.
ByteDance offers Seedance 2.5 to developers through BytePlus, and several API platforms resell it. This site is a web app for creating videos without code; it does not offer its own API.
No. Seedance25Video is an independent service that runs Seedance 2.5 through a third-party API provider. Seedance is a trademark of ByteDance.
Write a prompt or upload a photo, pick a model and get a clip with sound in a few minutes. Plans start at $14.90 a month, and credits are only spent on videos that finish.
Seedance is a trademark of ByteDance. Seedance25Video is an independent service and is not affiliated with or endorsed by ByteDance.
