A published boundary, not the absence of one
Uncensored AI video generator
Most AI video tools refuse ordinary creative premises. A horror beat, a courtroom insult, a character who dies at the end of the sketch, a joke with teeth, all get blocked at the filter. Frutti runs a narrow, published boundary instead of a broad one: it writes dark, absurd and uncomfortable premises without arguing with the writer, and lists, in plain language, the things it will not do.
Real outputs, not a promo reel
Finished, self-hosted Frutti episodes — the same engine and pipeline this generator runs. Judge the output before paying anything.
Write your first uncensored AI episode.
Pick a premise or write your own, and read the episode before you pay for anything. No prompt engineering, no editing, no account.
Story
Your one-line premise becomes a scene-by-scene script with a conflict and a payoff.
Cast
Original characters are designed and kept consistent from scene to scene.
Voices
Every character gets a distinct AI voice — dialogue is performed, not subtitled.
Captions
Word-timed captions are burned in, styled to stay readable with the sound off.
Export
One finished 9:16 episode for TikTok, Shorts and Reels. No editor, no export settings.
What 'uncensored' actually means here
In this category, 'uncensored' usually means a vague marketing claim and an even vaguer safety policy. The honest version is a written boundary that names what is refused and why, not a promise that anything goes. Frutti publishes its boundary on this page rather than burying it in a terms-of-service link, because the readers who search this query have been burned by vague policies before and can tell the difference between a real limit and a liability hedge.
The engine is built to take a premise a writer would consider ordinary: a horror beat where someone survives, a courtroom sketch with an actual insult in it, a crime short where the perpetrator gets caught, a character with a sharp opinion. These get refused by tools whose filter is tuned to block anything ambiguous, which produces a tool that writes only the safest third of the prompts it receives. Frutti writes them and renders them.
There is a real ceiling, and it is lower than the ceiling most readers imagine. Anything involving minors in a sexual context, graphic gore, dismemberment or torture, sexual deepfakes of real people, and instructions for self-harm is refused and cannot be unlocked on any plan. The boundary is narrower than what most tools publish, not wider, and it is enforced by the same engine that produces the horror, courtroom and crime formats.
Premises other tools refuse
These are the kind of one-line premises that get bounced at the filter on most AI video tools. Each one is a finished Frutti episode: script, cast, performed dialogue, captions, vertical export.
- a horror short where the final jump scare is the last thing anyone sees in the apartment
- a courtroom show where the defendant admits the crime on the stand and still demands a retrial
- a crime short where the victim turns out to be the architect of the whole setup
- a horror beat where the narrator is the monster, narrated in first person the whole time
- a plush toy cast staging an intervention about the owner's sleep schedule, with teeth
A boundary you can read
A page that claims no limits reads as an advertisement and gets discounted on contact. Frutti publishes what it refuses in the same place it sells the tool, because a writer who needs a dark premise delivered without a fight needs to know the refusal list before they start typing, not after the third blocked render.
The numbers, so you can check them
- What you download
- A 9:16 vertical video with the captions already burned in, ready to post without an editor.
- Episode length
- Episodes are built from long dialogue clips, each one continuous shot: Quick 2 clips of 10 seconds, Standard 3 of 10, Long 3 of 15, Epic 4 of 15. Longer runs come from a series of episodes.
- Watermark
- None. No export carries a watermark, on any plan, and there is no watermarked free tier to upgrade out of.
- What is free
- Writing a full episode (script, cast and scene list) on the home page and on every format page, with no account and no card. Rendering runs on credits. There is no free render tier and no watermarked tier.
- Price
- Starter $24/month or $7/month billed yearly, Creator $59/month or $18/month billed yearly. Yearly billing is 70% off and is charged once for twelve months. Prices in US dollars.
- What an episode costs
- 600 credits for a Standard episode, 1,050 for a Long episode. Starter includes 1,200 credits a month, which is 2 Standard episodes or about 30 single clips.
- Languages
- English plus 16 other languages.
- Ownership
- You own the output you generate. Paid plans cover commercial use, including client work, brand channels and paid advertising.
Verified 4 September 2026. Prices, credits and clip counts are read from the configuration that charges for them, not typed into this page. Pricing, refunds, terms. Machine-readable: /api/public/plans and /llms.txt.
Frequently asked questions
Is Frutti really uncensored?
It runs on a published boundary, not the absence of one. Dark, absurd and uncomfortable premises are written and rendered without the filter arguing with the writer. A short, explicit list of permanently refused content is stated on this page, because hiding it in a terms-of-service link is the move that gives this category a bad name.
What will Frutti refuse, exactly?
Anything involving minors in a sexual context, graphic gore, dismemberment or torture, sexual deepfakes of real people, and instructions for self-harm. These are refused and cannot be unlocked on any plan, and the same limits apply whether the output is a horror short, a courtroom sketch or a talking fruit drama.
Why is the boundary narrower than 'anything goes'?
Because the point of a real boundary is to protect the writer from a tool that disappears on them mid-project, not to advertise an anything-goes product that would be shut down by the first payment processor or platform upload. A narrower, honest boundary is more useful than a broad one that gets quietly enforced by a different team.
Do the model layer and the platform layer add limits on top?
Yes, and this page says so plainly. The video model has its own safeguards running on top of Frutti's, and the platform you upload to (TikTok, YouTube, Instagram) applies its own rules to whatever you render. A clean render does not guarantee a clean upload; that is a separate question and a separate refusal list.
Can I use the output commercially?
Paid plans cover commercial use, including client work, brand channels and paid advertising. You own the output you generate on a paid plan. Writing the full episode is free and needs no account; rendering the video is paid.