Nodes/SpicyAPI/Seedance 2.0 Fast Spicy · Image to Video
ComfyUI Node

Seedance 2.0 Fast Spicy · Image to Video

Uncensored, and audio costs nothing extra

By spicyapi-ai·Created 21 days ago·Updated 2 days ago· 2
Seedance 2.0 Fast Spicy · Image to Video
  • image_url
  • last_image_url
  • video
  • task id
◄prompt►
◄aspect_ratioadaptive►
◄resolution720p►
◄duration_seconds5►
◄generate_audiotrue►
◄seed0►

Fast Spicy is the odd one out in the 2.0 Fast line, and in a useful way. It's bytedance/seedance-2.0-fast-spicy/image-to-video: a still goes in, 4 to 15 seconds of uncensored video comes out, at 480p or 720p, from $0.081 per second.

Compared to the standard Fast image-to-video node, it drops the knob clutter - no bitrate_mode, no watermark, no return_last_frame - and keeps only what you actually need. Compared to the Spicy tier's reputation, it's the cheapest uncensored Seedance 2.0 you can get through this pack. (Mini Spicy is cheaper still at $0.0387/s, and 2.0 Spicy proper is $0.114/s.)

The short input list

image_url is an IMAGE socket - Load Image wires straight in. The pack uploads the still for you and, on this model, only the image is required; the field order in the schema literally puts image_url before prompt, which tells you the picture is doing the heavy lifting.

prompt is where the action goes: what happens, camera movement, mood. aspect_ratio defaults to adaptive, which follows the ratio of the image you connected - the sane default here, and the reason you can drop in a still without re-framing it. resolution gives you 480p or 720p and nothing above; if you need 1080p or 4K, that's the full Seedance 2.0 Spicy node, at a higher rate.

duration_seconds is an integer slider from 4 to 15 - a real slider, not the -1-and-a-list dropdown the non-spicy Fast nodes use. There's no smart-duration mode on the Spicy variants; you say how long, and that's the billed quantity. generate_audio is on by default and produces native synchronised sound, and on this model switching it off saves you nothing.

seed behaves like every node in the pack: fixed control-after-generate by default, so re-queueing a graph reuses the cached result instead of quietly billing you twice. Randomize it when you want another take.

That's the whole node. Eight inputs, only two of which are creative (the image and the prompt), really.

Outputs

video - a standard ComfyUI VIDEO, straight into Save Video - and task id, a string matching the entry in your SpicyAPI console. If a run ever misbehaves, that id is the entire support ticket.

What the Spicy tier does and doesn't promise

Uncensored here means the endpoint runs with a looser content policy than the standard Seedance 2.0 route. It doesn't mean the model is a different network, and there are no weights you can inspect or patch. You are trusting the provider's description of its own moderation, so test on the content you care about before you build a project around it. That's the trade for a closed model; the KB's framing on API nodes applies unchanged - the right tool for something you could never run locally, and inputs that leave your machine by design.

Install

ComfyUI Manager → search SpicyAPI → Install. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/spicyapi-ai/comfyui-spicyapi

Restart, then Settings > SpicyAPI and paste a key from spicyapi.ai/console/keys, or set SPICY_API_KEY on a headless machine. One more restart and the node menu is built from your account's catalogue. Nothing to pip install; nothing to download.

Getting it to run cleanly

The usual four, in order of how often they happen: no key set ("No SpicyAPI key is set"), a spend cap that stops the run before it charges (Settings > SpicyAPI > Max cost per run - set one, video gets expensive), an interrupt that doesn't cancel the server-side job (it finishes and is charged, result in your console for up to 14 days), and a 40004 when the option combination isn't served. A stale node menu means you need a restart, not a reinstall.

CategorySpicyAPI/Video/ByteDance

Inputs (8)

NameTypeDefaultDescription
image_urlIMAGEStart image URL to guide the video generation.
promptSTRINGDescribe the scene, action, camera movement, and mood for the video.
aspect_ratioCOMBOadaptiveThe aspect ratio of the generated video. Use adaptive to follow the input image.
resolutionCOMBO720pThe output video resolution.
duration_secondsINT54–15The duration of the generated video in seconds (4-15s).
generate_audioBOOLEANtrueWhether to generate native audio synchronized with the output video. Defaults to true.
seedINT00–2147483647Same seed and settings, same result. Set 'control after generate' to randomize to get a new result on every run; every run is billed.
last_image_urloptIMAGELast frame image URL for video continuation.

Outputs (2)

NameTypeDescription
videoVIDEO—
task idSTRINGThe SpicyAPI task id, for the console logs or support.