Papiboom Studio

简体中文

IMAGE TO VIDEO / REFERENCE IMAGES

Image to video with reference images,
fewer filters, same rules.

Start a video from your own images, keep a character or a shot, and see which models fit adult creative work. No local install; everything runs in the browser.

No subscription required / Pay-as-you-go credits

Neutral sample output from the Papiboom workspace. This page shows no adult imagery.
Direct answer /

You can turn your own images into video with a broader adult range, but not without review. Seedance 2.5, the main model, takes up to 30 reference images and is the most expressive, but weaker on detailed anatomy; Wan 3.0 takes up to 10. Requests are still reviewed, and every person shown must be an adult who consented, as the safety rules require.

Video models that start from your images.

Image-to-video works in two ways: first/last frame (your image becomes the opening or closing shot) or reference mode (images guide character, style, or scene). How far each model goes with adult themes differs; see the ratings below.

Model
Mode
References
Best for
Seedance 2.5
Reference or first/last frameUp to 30 images, 10 videos, 10 audioMain model: most references; weaker anatomical detail
Seedance 2.0
Reference or first/last frameUp to 9 images, 3 videos, 3 audioCinematic motion; weaker anatomical detail
Seedance 2.0 Mini
Reference or first/last frameUp to 9 images, 3 videos, 3 audioLow-cost drafts; NSFW results inconsistent
Wan 3.0
Reference or first/last frameUp to 10 images, 5 videos, 5 audioLong clips at a low per-second price; top NSFW rating
Happy Horse 1.1
Image referenceUp to 9 images; no video or audioCharacter motion from images
WAN 2.7
First/last frame1 start image, optional end frameKeep an exact opening shot

Seedance is the main video model here, with the strongest expressiveness (motion, camera, and atmosphere), but it is weaker at rendering detailed human anatomy, especially intimate areas.

Video model ratings, including NSFW capability.

How each video model scores on seven dimensions, out of 5, in half-point steps. Seedance is listed first as the main model.

Swipe the table sideways to see every rating.

ModelVisual qualityMotion naturalnessCamera workReference consistencySpeedValue for moneyNSFW capabilityNotes
Seedance 2.555553.52.54Strong visuals, motion and camera work; relatively lower value rating.
Seedance 2.05554.53.53.53.5Strong visual quality, motion and camera work.
Seedance 2.0 Fast4.54.54.54.5452.5High value rating; NSFW results are inconsistent.
Seedance 2.0 Mini44.54.54452.5High value rating; NSFW results are inconsistent.
Wan 3.04.554.54.53.555Strong motion, value and NSFW capability ratings.
Wan 3.0 Prime4.554.54.5555Higher speed rating than Wan 3.0, with the same other scores.
Happy Horse 1.14.533.532.523.5Good visual quality; speed and value ratings are lower.
WAN 2.7343443.54.5High NSFW rating, with good motion and reference consistency.
Kling 4.0554.55331Not truly NSFW-capable, though moderation is relatively relaxed.
Kling 3.0 Motion Control54.5××230.5Strong quality and motion; lower speed and NSFW ratings.
Vidu Q44.54.54433×Strong visuals and motion; NSFW is unsupported.
MiniMax H34443.53.543.5Relatively balanced ratings across dimensions.
WAN 2.5232×333.5Weaker visuals and camera work; listed reference capability is unsupported.
Happy Horse 1.0433.52.52.523.5Decent visual quality; lower speed and value ratings.
Grok Imagine453.5×4.551Strong motion, speed and value; limited NSFW capability.
Grok Imagine 1.5 Preview453.5×4.52.51Strong motion and speed; lower value than Grok Imagine.

Source: internal extensive testing and user feedback, reviewed November 12, 2026. Subjective ratings, not standardized benchmarks. Results vary by input and task; success is not guaranteed. × means the model does not support that capability.

From reference image to video, step by step.

Every step stays on one canvas, so the source image, prompt, settings, and result remain connected and can be repeated.

01

Check the source

Use images you own or may use. Anyone shown must be an adult who consented to this use.

02

Add image nodes

Upload or generate the images and connect them to a video node.

03

Choose the mode

First/last frame to keep the shot; reference mode to keep a character or style.

04

Write the direction

Describe action, camera, light, and pacing. Refer to each image by its name.

05

Check the quote

The credit quote appears before you submit; reference video adds billed seconds.

06

Review the result

A rejection shows "could not pass the safety review" and returns the credits. Revise within the rules.

Reference video pricing example.

On Seedance 2.5, a reference video bills a lower rate on output seconds plus reference seconds. A 5-second clip with a 5-second reference video costs 76 credits at 480p, 168 at 720p, and 420 at 1080p. Without a reference video the same clip costs 62, 140, and 350.

What "fewer restrictions" means here.

"Uncensored" often promises no rules at all. Papiboom offers a broader adult creative range, with content review and safety rules still in place.

01 / CREATIVE RANGE

Broader

Selected image and video models accept a wider range of adult themes than mainstream generators.

02 / CONTENT REVIEW

Still applies

Requests are still reviewed and some are rejected. A rejected task shows "could not pass the safety review" and its credits are returned.

03 / PLATFORM RULES

Always on

The adults-only, consent, real-person, and prohibited-content rules apply to every model and every account.

What stays in force on every model.

These rules come from the Papiboom Safety and Consent Policy. A model with less filtering does not change them.

Required

Before you create adult material.

  • You confirmed at sign-up that you are 18 or older
  • Everyone depicted is an adult and has consented to that use
  • You have the rights to every reference image, video, or voice
  • The work complies with the law where you live and publish

Prohibited

No model, label, or prompt makes these acceptable.

  • Anyone under 18, presented as under 18, or minor-looking
  • Sexual or intimate depictions of a real person without their express consent
  • Non-consensual intimate imagery, sexual deepfakes, coercion, or exploitation
  • Attempts to evade or test safeguards

Questions people ask before they start.

Short, factual answers. Prices and limits were checked against the live product on November 12, 2026.

Is there an uncensored image to video AI with reference images?

Papiboom offers image-to-video with reference images on models with a broader adult range, led by Seedance 2.5. It is not uncensored in the sense of no review: content review and the safety rules still apply.

Can I use NSFW images as references?

Adult material is allowed only within the safety rules: every person shown must be an adult who consented to that use, and you need the rights to the image. Real people without consent and anyone minor-looking are prohibited.

Do I need to install anything locally?

No. Papiboom runs in the browser; uploads, generation, and editing happen on the canvas.

How many reference images can I use?

Up to 30 on Seedance 2.5, 10 on Wan 3.0, and 9 on Seedance 2.0, Seedance 2.0 Mini, and Happy Horse 1.1.

What if my reference is rejected?

The task shows that it could not pass the safety review and its credits are returned. Revise the request within the rules; Papiboom does not provide ways around review.

Adults only. Rules first, then create.

Creating an account requires confirming that you are 18 or older and accepting the Terms. Read the safety and consent rules before you start.

Create account