Vidu —
AI video + no waitlist
Vidu (Shengshu) — video model with solid character consistency and reference support. See what it can do and how to get similar results through Viralio.
What Vidu is
Vidu (Shengshu) — video model with solid character consistency and reference support. It is part of the wave of models that made content generation available to anyone in 2026.
Direct access can be limited or waitlisted. On Viralio you get equivalent results with Veo 3 (native audio) and Seedream (smooth motion), in a studio in English, paying in euros — no international account or card.
Strengths
Prompt examples
You need no separate account and no waitlist. On Viralio you get Veo 3 (native audio) and Seedream (smooth motion) in a single studio — you generate in English, pay in euros and start with 10 free credits.
How to write a good prompt for Vidu (the full formula)
Nine times out of ten, the difference between a forgettable clip and one that pulls thousands of views is the prompt. An AI model does not guess what you want — it turns your words into image and motion. The more precisely you describe, the closer the result is to what is in your head.
The formula experienced creators use has seven ingredients, in this order: Subject → Action → Setting → Light → Camera move → Format → Audio/mood. You do not need all of them every time, but each extra element buys you extra control.
The 7 ingredients, explained
- The subject — who or what is in frame. Be concrete: "a young woman with red hair and a leather jacket", not just "a person".
- The action — what the subject is doing. Clear verbs: "pours", "runs", "turns slowly towards the camera".
- The setting — where the scene happens. "in a bohemian café with big windows", "on a wet city-centre street at night".
- The light — the most underrated element. "golden hour", "pink neon light", "soft morning window light", "dramatic backlight".
- The camera move — the standard cinematography terms work best: dolly in, pan, tilt, tracking shot, crane up, orbit, handheld.
- The format — 9:16 for TikTok/Reels/Shorts, 16:9 for YouTube and websites, 1:1 or 4:5 for the feed. Say it explicitly.
- Audio/mood — on models with native audio (Veo 3) you ask for the sound directly: "the sound of rain", "a voice saying…", or "complete silence". On the rest, describe the mood: "cinematic, melancholic".
Start from a simple idea — "a coffee" — and turn it into a prompt that comes out well on Vidu:
A barista pours latte art into a white cup (subject + action), on a wooden counter in a café with warm morning light (setting + light), macro with a rack focus from the steam to the cup (camera move), vertical 9:16 (format), the sound of steaming milk and background chatter (audio).
Notice how every added detail reduces the model's freedom to wander off. Beginners write five-word prompts and wonder why the results look generic; professionals write two or three dense sentences.
The full workflow: from idea to post
Here are the exact steps a creator goes through to produce post-ready content, with no team and no editing software:
- Start from an idea and a platformDecide what you want to say and where you are posting (TikTok, Reels, Shorts, YouTube, feed). The platform dictates the format and the length.
- Write the prompt using the formulaSubject + action + setting + light + camera move + format + audio. The more specific, the better.
- Pick the right modelVeo 3 for cinematic with native audio, Seedream for smooth vertical (TikTok/Reels), Grok Aurora for artistic. Or let the automatic recommendation choose.
- Generate 2-3 variantsDo not expect perfection first time. Change one element at a time (the light, the angle, the format) and compare. Each variant costs only a few credits.
- Add AI voice and subtitlesAI voice narration in English plus automatic subtitles. 85% of people watch with the sound off — subtitles lift both reach and retention.
- Publish and measurePost at your audience's peak hour, watch what works and double down on the winning format. Repeat consistently.
Common mistakes (and how to avoid them)
What it costs and how to save credits
On Viralio you pay for what you actually use, through credits. A clip or an image costs a few credits depending on the model and settings — premium models (native audio, high resolution) cost more per generation than the fast ones.
Three habits that make your credits go several times further:
- Write a good prompt first time — every failed variant is a credit spent for nothing. Put 30 seconds into the prompt and you save generations.
- Test cheap, finish expensive — use a fast model to nail the composition, then regenerate the final version on the premium model.
- Reuse — a good clip reframes for TikTok, Reels and Shorts; a good image becomes a banner, a post and a story. Do not regenerate what you can adapt.
- Vidu gives results in proportion to how specific your prompt is — use the 7-ingredient formula.
- On Viralio you use it in English, billed in euros, with every other model in the same subscription.
- Generate 2-3 variants, add AI voice and subtitles, then publish consistently.
- Start free with 10 credits — enough to get the hang of it.
Viralio pricing — access to every AI model
One subscription opens every app and every model on Viralio. Automatic invoice, cancel any time. Start free, no card.
Pay annually and save up to ~40%. Teams & companies: the Business plan (from €40/seat) with a shared credit pool. See all the prices and the full comparison →
Frequently asked questions
Generate in Vidu style
Instant access to every model, with an EU invoice.
10 credits on us, no card.