Uncensored video and image edit on one API key, with prices

#ai #api #video #tutorial

Two uncensored media models went live on UplinkSwitch on 2026-10-19: Wan 3.0 text to video with the safety filter off, billed per second, and Seedream 5.0 Lite Edit with the filter off, billed per image. They run on the same key and prepaid balance as the 11 uncensored chat models and every other model in the catalog.

The two models

Model	Id	Price	Input	Output
Wan 3.0 Video Uncensored	`wan-3.0-video-uncensored`	$0.15 a second at 480p, $0.30 at 720p, $0.60 at 1080p	text prompt	a clip of 2 to 30 seconds, audio on by default
Seedream 5.0 Lite Edit Uncensored	`seedream-5.0-lite-edit-uncensored`	$0.0313 an image	prompt and `image_urls`	up to 15 images a request, up to 3K

Prices from GET /v1/catalog/media on 2026-10-19. Each model has a filtered twin at the same price (wan-3.0-text-to-video, seedream-5.0-lite-edit). The only difference is enable_safety_checker: false on the uncensored id, true on the twin. Send true on a request and the filter is back.

What we got today

A 2-second 480p clip (854x480, H.264, 60 frames) came back in about three minutes and cost $0.30. An image made with Lustify SDXL Uncensored, edited into a 2048x2048 PNG, took about a minute and cost $0.0313. Both runs are on the model pages: video, edit.

The video request

curl https://api.uplinkswitch.com/v1/generations \
  -H "Authorization: Bearer $UPLINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan-3.0-video-uncensored",
    "input": {
      "prompt": "A red paper boat on a calm blue pond, morning light",
      "resolution": "480p",
      "duration": 2
    }
  }'

The answer comes back at once with status: queued and the charge already set. Poll GET /v1/generations/<id> until status is succeeded; the file is a URL on api.uplinkswitch.com, kept for 30 days.

The edit request

from openai import OpenAI
client = OpenAI(base_url="https://api.uplinkswitch.com/v1", api_key="sk-uplink-live-...")
result = client.images.generate(
    model="seedream-5.0-lite-edit-uncensored",
    prompt="Replace the sky with a storm",
    extra_body={"input": {"image_urls": ["https://api.uplinkswitch.com/v1/files/.../0.png"]}},
)
print(result.data[0].url)

image_urls is required. The same call also works on POST /v1/generations with the model and its input.

The two rules

No image or clip may depict a minor. A sexual edit runs only on an image generated on UplinkSwitch, a /v1/files URL from your own request, never on a photo of a real person. A request that breaks either rule is refused with 400 content_not_allowed before anything runs, so nothing is charged.

Billing

The price is fixed when the request is accepted and charged then. A generation that fails is refunded in full. Credit is prepaid from $10 with no subscription, and every model in the catalog gets the same daily test prompt, with the result public at uplinkswitch.ai/proof: 75 of 86 models stable on 2026-10-19.

All seven uncensored image, video and edit models, with the chat models, are on one page: uplinkswitch.ai/uncensored. The five uncensored image generators, from $0.003 an image, are measured in this post.

DEV Community