Veo 3.1 is Google's cinematic video model, live in the Picasso IA catalog, but whether it filters an adult prompt has not been tested live. Try it first: if it completes your scene, you have your video. If it shows a Content Not Supported by This Model message instead, the app carries your prompt straight over to Picasso Video or LTX 2.3 Pro, the two video models confirmed to run with no content filter. Adults only, fictional characters only.
Veo 3.1 is Google's video generation model, and on paper it is a strong one: cinematic camera motion, coherent scenes held across several seconds of footage, and prompt adherence that keeps a subject looking the same from the first frame to the last. Naturally, the question that follows is whether it will complete an adult scene or refuse it outright. This page does not pretend to know the answer before you have tried it. What it does is walk through what actually happens either way, why Google's broader pattern makes a filter the likelier outcome, and the two video models on Picasso IA already confirmed to carry none, so a refusal never has to be the end of the attempt.
Veo 3.1 is a real model, live in the Picasso IA catalog today, and there is nothing stopping you from pointing an adult prompt at it. Whether it completes that prompt or declines it is the honest open question this page will not paper over. Picasso IA has not run that specific test live yet, so nothing here asserts a flat "it will refuse" or a flat "it will comply." What can be said is the context around it: Veo 3.1 comes from Google, and Google's other generative products are known for applying strict content policies to sexual material as a matter of stated company practice. A model built inside that same company is a reasonable bet to inherit that stance. A bet is not a confirmation, though, and the only way to close that gap is to run the prompt and read what comes back.
That is the right way to approach it in practice. Open the toolkit, write the scene out in full, and run it on Veo 3.1 first. If the clip comes back, the question is answered and you have a video besides, one with the cinematic camera movement Veo 3.1 is built for. If it does not, the app tells you exactly why in the same screen and hands you a way forward without losing anything you typed.
Many models across the Picasso IA catalog, image and video alike, run a safety classifier ahead of generation. The classifier reads the prompt, and sometimes the output, and checks it against a short list of categories the model's provider has restricted. Adult content sits on that list for most flagship models, next to a small set of categories every provider treats the same way no matter the subject: a real person's likeness, a public figure, and anything involving a minor. This is not a rule Picasso IA adds on top of the model. It is the model's own behavior, inherited from how the company that built it trained and shipped it.
Google is a useful case study here because its pattern is already visible elsewhere. Gemini and Imagen, the company's other widely used generative tools, are built with explicit, publicly stated policies against sexual content, enforced at the model level rather than left to a separate moderation pass. Veo 3.1 shares a company and a safety review process with those tools, which is the whole basis for expecting it to filter an adult prompt the same way. It is a pattern-based expectation, not a test result: worth expecting, not worth assuming as fact until Veo 3.1 has actually been run against a real prompt and the outcome logged.
When a filter does catch a prompt, on Veo 3.1 or any other model, it does not return a blurred frame or a warning image. It returns nothing, and the app shows a Content Not Supported by This Model message in its place. Nothing about your account gets flagged for future generations; the refusal is specific to that one model, on that one run. The message then recommends a small set of models confirmed to carry no content filter and carries your prompt over automatically, so switching models never means retyping the scene from memory.
One thing worth saying plainly here: the fix is never to reword or disguise a prompt to slip past a filter. That approach does not work reliably, it burns a generation on a guess, and it is explicitly discouraged across Picasso IA. The sanctioned move, the one the app itself offers the moment a filter fires, is switching to a model that does not filter adult content in the first place. That is what the steps below walk through.
The sequence below holds whether Veo 3.1 completes your prompt on the first attempt or not. It treats a refusal as a normal step in the process, something to route around, not a dead end to work against.
These are the two video models Picasso IA confirms carry no content filter today, and they cover different starting points rather than different subject matter.
Picasso Video starts from a still image you already have. Give it a keeper frame, a photo generated on an image model or one you already own the rights to, and describe the motion you want: a slow push in, hair moving in a breeze, a turn toward the camera. It animates that specific frame rather than inventing a new scene, which makes it the model to reach for once a still is already right and only needs to move.
LTX 2.3 Pro works the other way, generating a clip straight from a text prompt with a cinematic finish, no starting image required. It is the model to reach for when there is no still yet, or when the scene calls for camera work a single frame cannot capture on its own, like a tracking shot or a slow reveal.
Neither applies a safety filter to adult content, and both are the same two models the app recommends automatically the moment a filtered model, Veo 3.1 included, declines a prompt. Trying Veo 3.1 first costs nothing but a moment; having these two waiting either way means that moment is never wasted.
This applies on every model, filtered or not, confirmed or untested, and it is worth stating plainly rather than leaving to fine print. Everything this page describes is for adults, eighteen and over, and every subject must be a fictional adult character. Never a real person's likeness, with or without a claimed resemblance. Never a public figure. Never a minor, or anyone made to look like one, in any style, photographic or animated. Prompts that cross these lines get refused by the models themselves and violate the platform's terms; no wording gets around that, and none should be attempted, on Veo 3.1 or on Picasso Video or LTX 2.3 Pro.
A second limit is practical rather than moral. Current video models, filtered or not, still get hands, faces in motion, and background objects wrong often enough that nothing should leave the toolkit without a frame-by-frame look. And a last limit is about you rather than the model: what you generate and where you share it are yours to answer for, under local law and under the rules of whatever platform you publish to. Legal to generate is not automatically the same as allowed to post everywhere.
Generations on Picasso IA are paid for with credits, and video costs more per run than a single image, since a clip takes more compute to produce than one frame. New accounts start with free credits, which is usually enough to test whether Veo 3.1 completes a given prompt and, if it does not, to try the same prompt on Picasso Video or LTX 2.3 Pro right after, without spending anything out of pocket on the first round.
Exact costs depend on the model and the settings chosen, and they are listed in full, kept current, on the pricing page rather than repeated here where a number would go stale. As a general habit, if a scene needs several rounds of refinement, it is worth settling the composition and motion idea in your head before spending video credits on a version you already suspect needs another pass.
It is not confirmed either way. Veo 3.1 is a real model in the Picasso IA catalog, but whether it applies a content filter to adult prompts has not been tested live. Given how Google's other generative tools handle sexual content, a filter is the likelier outcome, but the only honest answer is to try it yourself and see what comes back.
It is what the app shows in place of a result when a model's own content filter flags your prompt before generation completes. It means that specific model declined that specific prompt, nothing more; your account is not affected, and the message comes with a recommendation to switch to a model confirmed to carry no filter.
No. When the app recommends Picasso Video or LTX 2.3 Pro after a refusal, it carries the prompt over automatically, so you switch models and generate again without retyping the scene from memory.
Two: Picasso Video, which animates a still image you provide, and LTX 2.3 Pro, which generates a cinematic clip from a text prompt. These are the only video models on the platform currently confirmed filter free; every other video model, Veo 3.1 included, has not been confirmed either way.
Picasso Video needs a starting image and animates it, so it is the choice when you already have a still you want to bring to life. LTX 2.3 Pro generates video straight from a text prompt with a cinematic finish, so it is the choice when there is no image yet or the scene needs camera movement a single still cannot show on its own.
Because Veo 3.1 comes from the same company as Gemini and Imagen, Google's other widely used generative tools, both of which are built with explicit policies against sexual content. That shared origin is why a filter is the expected pattern for Veo 3.1 too, though it remains an expectation rather than a confirmed result until the prompt has actually been run and the outcome observed.
No. Real people's likenesses and public figures are off limits on every model on Picasso IA, filtered or not, confirmed or untested, and prompts attempting either are refused and violate the platform's terms. The characters described on this page are fictional adults only.
Yes. The content on this page is for adults, eighteen and over, and every subject in it must be a fictional adult character. Nothing involving a minor, or anyone made to look like one, is permitted in any style, and prompts that try are refused.
It is paid in credits, and video runs cost more than a single image because a clip takes more compute to produce. New accounts get free credits to start with, enough to try Veo 3.1 and then, if needed, Picasso Video or LTX 2.3 Pro on the same prompt. Current, exact pricing lives on the pricing page.
No. Picasso IA runs in a browser, on a phone or a laptop, and the models run on hosted infrastructure rather than your device. There is nothing to install and no local hardware requirement, for Veo 3.1 or any of the other 488 models in the full catalog.
