Prompt filters
AI image generation without a prompt filter
An AI image generator with no prompt filter is one where your text is not matched against a list of terms and not rewritten on the way through. Vellria works that way, with one narrow exception: your wording reaches the model as typed, on the models that ship a prompt rewriter as a setting it is sent switched off, and the only refusals in front of a run are the prohibited categories.
What a refusal is usually rejecting
A prompt filter reads your text before anything is rendered and compares it against a list of terms and patterns. When something matches, you get a refusal, and the message is almost always the same regardless of what triggered it. That is why the experience feels arbitrary. The filter is reacting to wording, so a scene that passes when described one way is stopped when described another, and you are left rewriting blind against a rule you cannot see.
The blindness is the expensive part, not the refusal. You cannot tell whether the problem was a noun, a verb, an adjective it read as an intent, or a combination that means nothing on its own. So you try synonyms. The list you are guessing against was tuned for a mass consumer product and for the complaints its operator expects, which has very little to do with what you were asking for.
Models
Vellria Lite Image
The cheapest image model in the catalog. Prompt only: it takes no reference image.
Text to image
1 credit
Seedream 5 Pro Image
The flagship image model, focused on detail and composition. 2K output only.
Text to image
8 credits
Seedream 5 Pro Image Edit
The flagship editing model. Takes up to ten references and preserves identity and lighting.
Image to image
8 credits
Qwen 3 Image
Strong on long, complex prompts; stays faithful to scene composition and text.
Text to image
5 credits
Qwen 3 Image Edit
Edits using the images you provide as references, preserving character and scene.
Image to image
5 credits
Qwen 3 Pro Image
The higher-priced Qwen tier. Prompt only, and every output shape costs the same.
Text to image
7 credits
Qwen 3 Pro Image Edit
The higher-priced Qwen editing tier. Images past the first add to what a run costs.
Image to image
7 credits
Seedream 5 Lite Image
Fast, cheap Seedream. Reaches 4K and is good at typography and poster work.
Text to image
5 credits
Seedream 5 Lite Image Edit
Takes up to fourteen references and redraws the images you provide.
Image to image
5 credits
Wan 2.6 Image
Prompt only, from a fixed list of output shapes that includes this family's widest.
Text to image
6 credits
Wan 2.6 Image Edit
Edits from up to four reference images, and none of them adds to what the run costs.
Image to image
6 creditsThe second layer, after the render
The other kind of block runs after the image exists. It is produced, a classifier looks at it, and the result is withheld from you. From where you sit the two are indistinguishable: a wait, then a message. They fail for different reasons and neither says which one happened.
This one has a bill attached. The model did the work, the compute was spent, and on a platform that charges per generation you have paid for a picture you were not allowed to see. It also fails on its own terms fairly often, because a classifier judging a whole image has less to go on than you had when you wrote the sentence.
The quieter failure: silent rewriting
A refusal at least announces itself. The harder problem is rewriting: some platforms expand or reword your prompt before it reaches the model, adding detail, tidying grammar, and smoothing whatever they read as risky. An image comes back, so nothing looks broken. It is simply not the image you described, and nothing tells you your text was edited in transit.
This is what makes prompting feel unreliable elsewhere. You change a word, the rewriter reacts to your change in its own way, and you end up debugging a sentence you never wrote. Some models in this catalog carry that rewriter as a switchable vendor setting and it is sent switched off; the rest have no rewriter to switch. Either way the text you submit is the text that is used.
That is a property of the request, not a preference stored on your account. It holds the same whether the run came from a model page or from your own code, because both build the same body and neither adds a layer the other does not have.
When nothing is filtered, your wording is the only variable
With nothing rewriting your prompt, the difference between two renders is what you wrote. That makes prompting more precise work rather than less. Describe where the subject sits in the frame, what the light is doing, what the lens is like. A stack of adjectives hands the composition back to the model, which is exactly the ambiguity you were trying to remove.
On most models one input is not yours to set, and it is worth knowing which: only some declare a seed, and where none is declared the run-to-run variation underneath your wording is not something you dial in. Read a change across a few runs rather than off a single pair, and change one thing at a time so that what moved between them is the edit you made. Against a rewriter the same experiment tells you nothing, because two variables moved and only one of them was yours.
This is also why a prompt worth keeping is worth saving verbatim. Nothing is inserted between the words and the render, so the sentence that worked is a record rather than an approximation. It carries sideways as well: every image endpoint in the catalog reads the same prompt field, so wording tuned on a cheap model runs on a heavier one without translation, and what differs between them is the render rather than your interpretation of some intermediate layer's habits.
What still gets refused, and why that is not a filter
Uncensored describes the prompt, not the policy. A subject filter guesses at intent from your phrasing and gets it wrong in both directions, blocking ordinary work and missing what it was built for. The refusals we apply are a category rather than a guess: anyone who is not of age, and real, recognisable people used without consent.
The difference matters in practice because a category does not need to read your tone. Nothing has to infer whether you meant something innocently, and nothing loosens or tightens with usage. What that category does not cover is the provider that runs the model, which keeps a filter of its own on the requests it receives, judges them on its terms rather than on the ones published here, and can stop a run once it has started. A refusal from there is recorded as its own kind of failure and the credits taken at the start are returned automatically, so what it costs you is a re-run rather than a balance. The page on what is allowed and what is blocked sets out the full policy, what happens to an account that crosses it, and who can see the work you generate.
Frequently asked questions
Why was my prompt refused by another AI image generator?
Most refusals come from a text filter matching your wording against a list of terms before anything is rendered, so the trigger is often the phrasing rather than the subject. Some tools also screen the finished image and withhold it. Both paths return the same generic message, which is why rewriting your way out of one feels like guesswork.
Does Vellria rewrite my prompt before rendering?
No. Where a model exposes a prompt rewriter as a vendor setting, the layer that expands or softens your text before generation, it is sent switched off; the models without that setting have nothing to rewrite with. The wording you submit is the wording the model receives.
Can I re-run a prompt and get exactly the same picture back?
No, and that is worth planning around. Only some models declare a seed, and even there the same number does not guarantee the same frame, so identical wording is a fresh render rather than a replay of an earlier one. Keep the prompt because it is a faithful record of what you asked for, and download the frame itself if that particular file is the one you need.
Will a prompt I tuned on one model work on another?
Yes. Every image endpoint in the catalog reads the same prompt field, so the wording moves across unchanged and you are comparing models rather than rewriting for each one. What differs between them is the render and the options each accepts, not the language you have to speak.
If nothing filters the prompt, is anything blocked at all?
Yes, and our own list is deliberately narrow: depictions of anyone who is not of age, and real people in non-consensual content, refused as a category and not softening with time or usage. Separately, the provider that runs the model filters the requests it receives on its own terms, and can stop a run after it starts; that one is refunded automatically. The page on what is allowed and what is blocked has the full list and the consequences.