Seedance 2.0
prospolabs/seedance-2
Seedance 2.0 with improved fidelity up to 1080p. 4-15s duration with always-on audio. Reference mode accepts up to 9 images, 3 videos, and 3 audio clips with native audio and cinematic camera control — reference them in the prompt as @Image1 / @Video1 / @Audio1.
Starts from
Retail $0.150
Input
Typically completes in 30–90s.
Result
IdleOr write your own prompt and hit Generate.
01overview
What is Seedance 2.0?
Seedance 2.0 is ByteDance's unified multimodal video model: it takes text, images, audio and video as input and returns cinematic video with native, always-on audio. Clips run 4 to 15 seconds at up to 1080p. On Prospolabs you call the exact same model the web UI uses, priced per second in plain USD with no tokens and no subscription.
What sets Seedance 2.0 apart is compositional control. Its reference mode composes up to 9 images, 3 videos and 3 audio clips that you address directly in the prompt as @Image1, @Video1 and @Audio1 — so you can pin a character, a location, a motion reference and a voice into a single shot. Audio and video are generated jointly, giving you lip-sync and stereo sound without a separate post-production pass.
Prospolabs runs Seedance 2.0 at $0.09/sec (480p), $0.18/sec (720p) and $0.41/sec (1080p) — each roughly 40% under the retail rate, with audio included and no charge on failed generations. Need lower latency and a cheaper bill, you can route the same prompt to Seedance 2.0 Fast; need cinematic lip-sync or pure motion work, compare Veo 3.1 and Kling V3. All of them run on the one Prospolabs API at one transparent price.
The cheapest reliable Seedance 2.0 API — with a real number, not vague "value"
Prospolabs publishes one transparent per-second table: 480p $0.09 (retail $0.15), 720p $0.18 (retail $0.30), 1080p $0.41 (retail $0.683) — roughly 40% off across every tier, audio included, no charge on failed runs. A 5-second 720p clip costs $0.90 instead of $1.50, and we list 1080p where most providers hide it. Pay per generation in plain USD, top up from $5, no tokens and no subscription. Need it cheaper and faster, route the same prompt to Seedance 2.0 Fast.
02capabilities
What Seedance 2.0 can do
Native joint audio + video
Audio is generated in the same pass as the picture, so dialogue lip-syncs and ambient sound, music and stereo land in-frame. There is no separate audio step and no extra charge — audio is always on across every resolution.
Reference mode with @-mention control
Compose up to 9 reference images, 3 videos and 3 audio clips (12 files max combined) and address them in the prompt as @Image1 / @Video1 / @Audio1. Pin a character, a location, a camera move and a voice into one shot for repeatable, directed output.
Three input modes, one model
Text-to-video for a clean prompt, image-to-video to animate a starting frame (with an optional end frame), and reference-to-video for full compositional control. Switch with a single mode field — same model id, same per-second price.
Up to 1080p, 4 to 15 seconds
Generate at 480p, 720p or 1080p with durations from 4 to 15 seconds and aspect ratios including 16:9 and 9:16. Prospolabs publishes 1080p pricing openly where most providers stop at 720p.
Cinematic camera control
Direct dolly, pan, orbit and fly-over moves from the prompt. Seedance 2.0 holds subject and scene consistency across the shot, which makes it well suited to directed, multi-beat sequences.
Pay-per-generation in USD
Priced per second of output — $0.09 (480p), $0.18 (720p), $0.41 (1080p). Top up from $5, no monthly tier, no token math, and failed runs are auto-refunded so you only pay for video you actually get.
03use cases
What people build with Seedance 2.0
Advertising and UGC at volume
Spin up product spots and social-native UGC with synced voiceover and music baked in. At $0.18/sec a 5-second 720p clip with audio is $0.90 versus $1.50 retail, so iterating ten variants stays cheap.
Film pre-viz and previsualization
Block out shots with cinematic camera moves and reference your concept art via @Image1 before committing to a shoot. 1080p output at $0.41/sec gives directors a watchable previz reel without a render farm.
Character-consistent series content
Use reference mode to pin a recurring character and voice across episodes — @Image1 for the face, @Audio1 for the voice — so the same persona shows up shot after shot.
Architecture and product walkthroughs
Generate fly-throughs of a building or a product from stills, with ambient sound. Image-to-video animates a single render into a moving sequence; reference mode blends multiple angles into one coherent move.
Gaming and fashion concepting
Prototype animated game cinematics or runway-style fashion clips fast. Start at 480p for cheap exploration, then re-run the winning prompt at 1080p when you are ready to finish.
04api
Call Seedance 2.0 from your code
Seedance 2.0 uses an async submit-poll-download flow on the Prospolabs API. You POST a prompt, receive a generation_id, poll until the job completes, then download the output_url (valid for 7 days). Same USD price for API and UI calls; failed runs are auto-refunded.
Full Seedance 2.0 API reference →- 01
POST to /v1/generate with model "seedance-2", a "mode" of text, image or reference, your prompt, plus resolution and duration. In reference mode, attach image_urls / video_urls / audio_urls and cite them in the prompt as @Image1, @Video1, @Audio1.
- 02
Read generation_id from the response — the job runs asynchronously, so the initial call returns immediately rather than the finished video.
- 03
Poll the job status by generation_id on an interval (start around 5 seconds) and back off to avoid hammering the endpoint while longer or higher-resolution clips render.
- 04
When status is complete, read output_url and download the MP4 within 7 days; the file includes the native audio track with no separate fetch.
- 05
Handle failures with a retry-and-backoff loop. Failed generations are auto-refunded on Prospolabs, so a retry costs you nothing for the run that errored.
?questions
Seedance 2.0 FAQ
On Prospolabs, Seedance 2.0 is priced per second of output: $0.09/sec at 480p, $0.18/sec at 720p and $0.41/sec at 1080p. Audio is always included at no extra charge. A 5-second 720p clip is $0.90; a 5-second 1080p clip is $2.05. You pay per generation in USD with no subscription or tokens.
Prospolabs runs Seedance 2.0 at $0.09/sec (480p), $0.18/sec (720p) and $0.41/sec (1080p) — about 40% below the retail rates of $0.15, $0.30 and $0.683. Audio is included, failed runs are auto-refunded, and you can start from a $5 top-up with no monthly fee. For an even lower bill, the Seedance 2.0 Fast tier costs less and runs at lower latency.
Yes. Seedance 2.0 generates audio and video together in a single pass, producing lip-synced dialogue, stereo sound, music and ambience with no separate post-production step. Audio is always on and is included in the Prospolabs per-second price at every resolution.
Standard Seedance 2.0 targets the best fidelity up to 1080p with full reference and audio support. Seedance 2.0 Fast is the cost-optimized, lower-latency tier for quick iteration and high-volume work at a lower per-second price. Both run on the same Prospolabs API; you choose based on whether you need top quality or faster, cheaper turnaround.
Reference mode accepts up to 9 images, 3 videos and 3 audio clips, with a combined limit of 12 files. You address each one in the prompt as @Image1, @Video1 and @Audio1 to direct composition. Audio references require at least one image or video to also be supplied.
No. Failed generations are auto-refunded on Prospolabs, so you only pay for video you actually receive. This makes a retry-with-backoff loop safe — the errored run costs you nothing.
Seedance 2.0 supports 480p, 720p and 1080p output, with clip durations from 4 to 15 seconds and aspect ratios including 16:9 and 9:16. On Prospolabs, pricing scales by the second at each resolution: $0.09 (480p), $0.18 (720p) and $0.41 (1080p).
They are specialized tools. Seedance 2.0 is the multimodal reference director — strong on compositional control via @-mention references and native joint audio. Veo 3.1 leans into cinematic realism and lip-sync, and Kling V3 is a motion specialist. On Prospolabs you can run all three through one API at transparent per-second prices and pick the right one per shot.
related