How to make Italian brainrot AI videos: character, prompts, voice and cost

This page is a workflow and a cost model: one original character, ready-to-paste prompts, and list-price arithmetic as of 2026-10-26. Change the retry counts to match your luck.

What is AI Italian brainrot?

AI Italian brainrot is a meme genre of absurd AI-generated hybrid creatures with Italian-sounding names, shown in short videos with Italian narration.

Wikipedia says the content is made with generative AI for images, video and Italian audio, and that Tralalero Tralala, a three-legged shark in sneakers, is widely considered the first example. A shark-in-shoes post using the audio, dated 13 January 2025, reached 7 million views and was later deleted. Euronews describes the voice as a deep computer-generated male voice singing in Italian, "or, at least, some Italian". Euronews quotes an animator in the genre saying "it's funny because it's nonsense", and that viewers mostly do not care how the images relate to the words.

That last point is useful: the genre runs on nonsense, so a strange clip can still work.

What do you need to make an Italian brainrot video?

An Italian brainrot video needs four parts: an original character design, a still image, a short animated clip and a voice line.

Wikipedia and Euronews do not identify the original creators' complete toolchains, so this page uses a stack that is easy to price. All of these models are on one API, Runway's, which makes the sums comparable.

One Runway credit is $0.01. Runway's models page does not say whether nano_banana_2.1 accepts reference images, so the plan below uses a text prompt and a single saved still.

How do you make an original brainrot character?

Make an original brainrot character by combining one everyday object with one animal or body part, then giving it a made-up Italian-sounding name.

Existing characters use this pattern, such as a shark with sneakers or a log with a face. Copy the pattern and invent your own character. The example here is a coffee cup on small legs, as in the hero image above. The name is invented for this page. Search TikTok for your own name before you publish, so it does not collide with an existing character.

Character sheet

• Name: Tazzino Pedalini
• Body: a white espresso cup with a handle, two short ceramic legs, red canvas sneakers
• Face: two small black dot eyes and a flat mouth on the cup's side
• Voice: slow, deep, dramatic male narrator speaking Italian-style words
• Line: "Tazzino Pedalini, caffè nelle gambe, corre corre corre."

Keep the same description words in every prompt. Changed wording gives a changed character.

What prompts make the image and the video?

Use one image prompt to fix the character, then three short video prompts that each move only one thing.

Image prompt

A glossy 3D render of a white espresso cup with two short ceramic legs and red canvas sneakers, two small black dot eyes and a flat mouth on its side, standing on a tiled Italian piazza at golden hour, low camera, stone fountain behind, vertical 9:16.

Generate several stills and keep the one that matches the sheet. To keep the subject steady when you animate it, see the Kling image to video settings.

Video prompts, using the still as the first frame

1. Shot 1: "The cup stands still, then blinks once and turns its eyes toward the camera. Slow push-in. Piazza light stays warm."
2. Shot 2: "The cup runs forward on its small legs, steam trailing from the top, camera tracking at ground level."
3. Shot 3: "The cup stops, tips slightly, and the espresso sloshes in the cup without spilling. Camera orbits a quarter turn."

Keep each clip at 5 seconds. Short clips cost less per retry. For more prompt wording, see the AI video prompt generator.

How do you add the Italian voice?

Add the voice by writing one short line and sending it to a text-to-speech model, or by letting a video model generate speech in the clip.

Runway lists Veo 3.1 as generating "audio and speech", and Kling 3.0 Standard audio as optional. For a separate voice, Runway's eleven_v3 model costs 1 credit per 50 characters. The line above is about 55 characters, so it costs 2 credits. A standalone brainrot voice site, Motionize, offers a preset voice at 10 credits per clip with a 1,000 character limit (as of 2026-10-26).

Keep real people's names, faces and cloned voices out of the line, because realistic content falls under stricter platform rules (see below).

How much does one Italian brainrot video cost?

One 15-second brainrot video with three 5-second clips costs between $4.31 and $13.76 at Runway list prices, once you plan for retries.

The assumptions: 5 image attempts, 3 clips with 3 attempts each, and 3 attempts at the one voice line. Change them to match your luck. A 55-character voice line is 1.1 credits, which I round up to 2 per try.

Veo 3.1 Fast only accepts 4, 6 or 8 second clips, so the Veo row prices six-second generations; trim each to 5 seconds in the edit. Seedance 2.5 also charges $0.15 per second of input video, which this plan does not use, and has an 80-credit minimum per generation. Its 150-credit clips are above that minimum. For per-model detail see the AI video cost calculator.

How do you edit the clips into a brainrot video?

Edit by cutting the three clips together, laying the voice line over them and adding large captions.

1. Trim each clip to its best 2 to 4 seconds. A 15-second video can use 10 seconds of footage.
2. Place the voice track across all three clips, not one per clip.
3. Add a loud, simple background beat or a sound effect on each cut.
4. Add captions.
5. Export vertical 9:16.

Any editor works; the plan leaves editing out of the cost.

Is it legal to make Italian brainrot videos?

Making original brainrot characters is the safe route, because ownership of existing ones is disputed and AI copyright law is unsettled.

Wikipedia says the copyright status of AI-generated content is an area of active litigation. It describes a dispute between Mementum Lab, which represents the original prompters and argues AI images can be copyrighted, and Do Big Studios, maker of the Roblox game Steal a Brainrot, which argues Italian brainrot is public domain. The dispute centers on Tung Tung Tung Sahur.

Platform rules apply too. YouTube's help page says disclosure is needed for realistic content that makes a real person appear to say or do something they did not, but exempts non-realistic content such as a fantastical animated scene. TikTok's guidelines, effective 24 September 2026, require labels on AI content with realistic-looking scenes or people, and say unlabeled content may be removed, restricted or labeled. A talking coffee cup is unrealistic. A realistic AI version of a real celebrity's face or voice needs a label in any trend; YouTube exempts cloning your own voice for voiceovers or dubs. Euronews also reports a Bombardiro Crocodilo clip that drew outrage for seemingly mocking the war in Gaza, so check what your audio says before posting.

What do the top brainrot generator pages leave out?

Motionize and Overchat sell a voice or a character generator and give no image prompts, video settings or cost per video.

As of 2026-10-26, Motionize's page describes a narrator voice and says you can send audio to a video or lip sync tool, with no steps for the visuals. Overchat's page produces a character name, backstory and image, and does not describe making video. Overchat does not name its model and states both a one-per-day free limit and "no limits". Wikipedia and Euronews explain the genre well but do not identify the creators' toolchains. This page fills that gap with a priced, repeatable plan. Run one cheap Kling pass first, then scale the budget. For more formats like this, browse the viral AI video guides. If you want to see the same method applied to a longer series, read the AI Fruit Love Island guide.