With Seedance itself, this key cannot give you Seedance 2 unfiltered. The model ships from its developer with a filter on, and the Seedance rows here have no switch that turns it off. What the key offers instead is a group of uncensored video builds, reached through the same endpoint and drawn from the same balance. Below, both sets sit in one place: every row priced per second from the live catalog, R2V explained, and the rows that accept reference images named.
For the per-clip arithmetic and the request loop, read the uncensored video generator API post; the media page lists every uncensored image and video model. Venice's Seedance page, which came up in the 27 October 2026 results, lives at venice.ai, not here.
Who else answers this search, and what the answer is here
On 27 October 2026, Google's first result was a Reddit thread hunting for an unfiltered Seedance 2 service. Under it: vendor pages from Kenerate, Wiro, ZenCreator, VideoAny and Apipod, each with "uncensored Seedance" in the title, the Venice page for Seedance 2.0, and one post from Sogni. Four ads sat above all of it.
The short answer for this key: all four Seedance rows, 2.5, 2, 2 Fast and 2 Mini, are on the catalog, called by API and billed per second of video. None of their rows has an enable_safety_checker setting, so there is nothing to send; every clip passes through the developer's filter. Uncensored video here comes from a separate set of models, in the second table below, and those take the same request body. Both sets are priced from the live catalog on this page, so you can compare the two routes before you spend anything.
The Seedance rows: filtered, per second
These are the Seedance rows returned by GET /v1/catalog/media on 2026-11-16. Each is priced per second of finished video, and the rate changes with resolution. The final column repeats what the row itself says: the default of its safety-checker setting if it has one, or a note that it has none.
| Model | Settings | Per second | 5 s clip | 10 s clip | Filter, as the row states it |
|---|---|---|---|---|---|
| Seedance 2 | resolution 480p | $0.125 | $0.63 | $1.25 | no enable_safety_checker setting on the row |
| resolution 720p | $0.25 | $1.25 | $2.50 | ||
| resolution 1080p | $0.5625 | $2.81 | $5.63 | ||
| resolution 4k | $1.25 | $6.25 | $12.50 | ||
| Seedance 2 Fast | resolution 480p | $0.0875 | $0.44 | $0.88 | no enable_safety_checker setting on the row |
| resolution 720p | $0.175 | $0.88 | $1.75 | ||
| Seedance 2 Mini | resolution 480p | $0.0625 | $0.31 | $0.63 | no enable_safety_checker setting on the row |
| resolution 720p | $0.15 | $0.75 | $1.50 | ||
| Seedance 2.5 | resolution 480p | $0.56 | $2.80 | $5.60 | no enable_safety_checker setting on the row |
| resolution 720p | $1.26 | $6.30 | $12.60 | ||
| resolution 1080p | $2.28 | $11.40 | $22.80 |
Seedance 2 accepts image_urls, reference_image_urls, reference_audio_urls plus a generate_audio switch, at resolutions of 480p, 720p, 1080p, 4k. If the developer's filter turns a prompt down, the generation comes back failed and the charge is reversed, which is the refund rule for every media model on this key.
If you searched "uncensored seedance", here is the plain fact: a Seedance clip from this key is filtered, and no setting on the key removes that. What you can get instead is a separate group of uncensored video builds, each listed under its own name, in the table that follows.
The uncensored video rows, per second
Each of these rows is an uncensored build of a video model. Text alone drives wan-3.0-video-uncensored; the two image-to-video rows, ltx-2.3-i2v-uncensored and wan-3.0-i2v-uncensored, start from an image as the first frame, which turns a still from an uncensored image model into motion. The filter column once more quotes the row: the checker's default when the setting exists, and when an uncensored build has no such setting, a note saying so.
| Model | Settings | Per second | 5 s clip | 10 s clip | Filter, as the row states it |
|---|---|---|---|---|---|
| Wan 3.0 Video Uncensored | resolution 480p | $0.20 | $1.00 | $2.00 | enable_safety_checker defaults to false |
| resolution 720p | $0.40 | $2.00 | $4.00 | ||
| resolution 1080p | $0.80 | $4.00 | $8.00 | ||
| Wan 3.0 Image-to-Video Uncensored | resolution 480p | $0.20 | $1.00 | $2.00 | uncensored build with no checker setting |
| resolution 720p | $0.40 | $2.00 | $4.00 | ||
| resolution 1080p | $0.80 | $4.00 | $8.00 | ||
| LTX-2.3 Image-to-Video Uncensored | resolution 480p | $0.08 | $0.40 | $0.80 | uncensored build with no checker setting |
| resolution 720p | $0.16 | $0.80 | $1.60 | ||
| resolution 1080p | $0.24 | $1.20 | $2.40 |
You do not need a terminal for any of this. The same rows run in the console at https://app.unrestrictable.ai/console/media, and the uncensored video generator page explains that path. By API, the uncensored video generator API post prices clips of 5, 10 and 15 seconds, walks through submit, poll and download, and totals up 100 clips. A single request:
curl https://api.unrestrictable.ai/v1/generations \ -H "Authorization: Bearer sk-unr-live-…" \ -H "Content-Type: application/json" \ -d '{ "model": "wan-3.0-video-uncensored", "input": { "prompt": "a dancer on a rooftop at dusk, slow dolly in, city lights behind", "resolution": "480p", "duration": 5 } }'The 202 response states the charge, seconds times rate, before any frame exists. Afterwards, ffprobe will print the file's duration so you can match it to the seconds billed:
ffprobe -v error -show_entries format=duration -of default=nw=1:nk=1 clip.mp4R2V on Seedance 2.0, in plain terms
R2V stands for reference-to-video. Rather than a single opening frame, you supply reference pictures of a character, a prop or a location, and write a prompt for a new shot; the model holds those references steady across the clip. Image-to-video (I2V) takes one picture and animates it from the first frame. Text-to-video (T2V) uses no picture.
On this key, references travel in reference_image_urls, while image_urls carries the first frame. Here is what each row on this page accepts as of 2026-11-16:
| Model | Picture and audio inputs | Lowest rate, per second |
|---|---|---|
| Seedance 2 | image_urls, reference_image_urls, reference_audio_urls | $0.125 |
| Seedance 2 Fast | image_urls, reference_image_urls, reference_audio_urls | $0.0875 |
| Seedance 2 Mini | image_urls, reference_image_urls, reference_audio_urls | $0.0625 |
| Seedance 2.5 | image_urls, reference_image_urls, reference_audio_urls | $0.56 |
| Wan 3.0 Video Uncensored | text prompt only | $0.20 |
| Wan 3.0 Image-to-Video Uncensored | image_urls (required) | $0.20 |
| LTX-2.3 Image-to-Video Uncensored | image_urls (required) | $0.08 |
{ "model": "seedance-2", "input": { "prompt": "the same character walks through a rain-lit market", "reference_image_urls": ["https://example.com/character.png"], "resolution": "720p", "duration": 5 }}The uncensored image-to-video rows accept image_urls and nothing more: one first frame, no reference set. A consistent character on the uncensored route therefore starts with an image model and finishes with I2V. Generate the still on one of the uncensored image models from the image and video page, then hand its URL to ltx-2.3-i2v-uncensored or wan-3.0-i2v-uncensored in image_urls.
The full uncensored media list on this key
The video rows from above together with the uncensored image models that feed them, every one on POST /v1/generations and priced per image or per second. For the image prices in detail, see the uncensored image generator API post.
| Per | Endpoints | |||
|---|---|---|---|---|
| LTX-2.3 Image-to-Video Uncensored | Video | $0.08 | second | |
| Wan 3.0 Image-to-Video Uncensored | Video | $0.20 | second | |
| Wan 3.0 Video Uncensored | Video | $0.20 | second | |
| Anime Image Uncensored | Images | $0.03 | image | |
| Chroma Uncensored | Images | $0.08 | image | |
| CyberRealistic XL Uncensored | Images | $0.0136 | image | |
| Qwen Edit Uncensored | Images | $0.08 | image | |
| Qwen Image 2.1 Uncensored | Images | $0.0375 | image | |
| Qwen Image 3 Uncensored | Images | $0.08 | image | |
| Seedream 5.0 Lite Edit Uncensored | Images | $0.0313 | image | |
| Vision Image Uncensored | Images | $0.04 | image | |
| Vision Pro Image Uncensored | Images | $0.09 | image | |
| WAI Illustrious SDXL Uncensored | Images | $0.0136 | image |
Questions
Not from this key. Seedance 2 is served by its developer behind a content filter, there are no public weights, and the rows here have no safety-checker setting to pass. For uncensored video, this key offers different models: the text-to-video build wan-3.0-video-uncensored plus two image-to-video builds, with per-second prices in the table. See the image and video page.
There is not. Seedance 2 has no public weights, so anything offered as a download is either a web app or some other model. The uncensored video builds here run over the API and need no GPU of your own.
Seedance rows pass through the developer’s filter, and a declined clip has its charge reversed. The uncensored video rows are their own builds; the table lists each row’s checker setting. Across all uncensored image and video models, a prompt that describes a minor is turned down before any charge, and the image-to-video rows accept only images that were generated here. The acceptable use policy spells out the rest. See the acceptable use policy.
Reference-to-video. You send reference images (and, on Seedance 2, reference audio) with the prompt, and the model keeps them consistent for the length of the clip. The field on this key is reference_image_urls; the table above marks the rows on this page that accept it.
Not on this key. You prepay credit from $10, each clip is priced per second before rendering starts, and a failed clip is refunded in full. See what 100 clips cost.
Venice has a Seedance 2.0 page that showed up in the 27 October 2026 results. Its claims are its own; read them at venice.ai. On this key, Seedance 2 keeps the developer's filter, and uncensored clips come from the uncensored rows listed above.
Every model above is on the uncensored model list with its live price per million tokens. Per-token pricing sets the whole catalog side by side, and the family directory groups the builds by base model.