Reeldo AIReeldo AI

Vidu Q4 AI Video Generator

Reference-led clips with native audio — up to 15 images, 3 voice samples, 3–16 seconds, and 540p–4K output. Image-to-video and reference-to-video only (via Kie on Reeldo).

  • 15 image refs
  • 3 voice refs
  • Native audio
  • Up to 4K · 16s

Reeldo is not affiliated with Shengshu Technology / Vidu. Specs follow Kie API docs.

Generate with Vidu Q4 Preview

No text-to-video on Q4 — use Image to Video for one strong frame, or Reference to Video for characters, scenes, and voice refs. Write dialogue in the prompt; MP3 refs set voice identity only.

Animate one start frame — output keeps the image aspect ratio. Toggle native audio on or off.

See Vidu Q4 examples below

Model

Need a start frame? Generate in Frame Lab

Start Frame

Required — the image Reeldo will animate from

End Frame

Optional — guide how the clip should finish

0/5000
16:9 · 5s · 720p · Example 1

Multimodal references, cinematic motion, and built-in audio in one generation.

Vidu Q4 example prompts

Copy a template or click Use in generator on Example 1. Reference audio carries voice identity — not the spoken words.

R2V · 1080p · 12s

Short drama beat with voice ref

Scenario
You have character + location stills and a clean voice sample, need one dialogue shot with lip sync.
Result
Reference images lock look; reference audio 1 guides timbre; prompt supplies new lines.
Settings

Prompt template

In the scene of reference image 1 (cozy workshop), the woman from reference image 2 turns to camera. She says: "You came back." in the voice from reference audio 1. Slow push-in, warm practical lighting, native room tone. Single continuous take.

Concept visual — upload your own refs in the generator.

I2V · 9:16 · 8s

Product hero for Reels

Scenario
You have one approved packshot and need vertical motion with subtle sound.
Result
Start frame locked; slow arc move; optional native ambience.
Settings

Prompt template

Vertical product hero from the start frame. Slow 8-second arc orbit — keep label and packaging shape unchanged. Soft studio light, subtle reflection, quiet native ambience (no music bed).

Concept visual — upload your own refs in the generator.

Marketing visual: one product packshot animates into a vertical Reels orbit hero with ambient waveform

R2V · voice ref · 5s

Talking head with fixed voice

Scenario
You need the same vocal identity across clips with new scripts.
Result
Voice ref sets timbre; prompt writes the new hook line.
Settings

Prompt template

The creator from reference image 1 speaks to camera in a bright home office. They say: "Here is what changed this week." in the voice from reference audio 1. Medium close-up, steady framing, native audio on.

Concept visual — upload your own refs in the generator.

Marketing visual: creator reference plus voice wave drives lip-sync talking-head frames

I2V · silent · 8s

Cinematic move without dialogue

Scenario
You want motion proof before paying for 4K audio-on finals.
Result
720p silent pass to lock camera path.
Settings

Prompt template

Cinematic 16:9 from the start frame. Slow crane up over 8 seconds — fog rolls through the valley, golden hour light. No dialogue. Turn native audio off for a silent proof.

Concept visual — upload your own refs in the generator.

Marketing visual: single landscape start frame and filmstrip showing crane-up motion path

How Vidu Q4 voice references work

MP3 voice refs (3–12 seconds each, up to 3) set vocal identity and delivery — they do not replace the script.

Diagram: MP3 voice reference plus prompt text drives lip-sync output on Vidu Q4
Reference imageVoice ref (MP3)Dialogue in promptLip-sync clip
Common belief

The uploaded MP3 is the dialogue

Reality on Vidu Q4

Wrong — write every spoken line in the prompt and point it at reference audio 1/2/3 for timbre.

Common belief

The model repeats the recording

Reality on Vidu Q4

It uses the recording as a voice example; new words must be in the prompt.

Common belief

Lip sync is guaranteed

Reality on Vidu Q4

Review mouth timing, word completeness, and emotion separately — iterate one variable at a time.

Pick your Vidu Q4 path

Q4 is reference-driven — start from assets, not a blank text prompt.

Vidu Q4 image-to-video from one hero frame

One strong frame

You only have a hero still.

Image to video

Vidu Q4 reference-to-video with character scene and voice ref

Character + scene + voice

Identity and timbre must stay aligned.

Reference to video

Text-only ideas route to other Reeldo text-to-video models

Text-only idea

Q4 has no text-to-video.

Use /text-to-video (Seedance, Kling, H3…)

Extend clips longer than Vidu Q4 sixteen second cap

Need >16 seconds

Q4 caps at 16s per job.

Video Extend or multi-clip edit

Vidu Q4 credits (per output second)

A Credits ≈ output seconds × rate at your resolution (540p–4K).

B Vidu Q4 reference mode uses the same per-second rate — no reference-video billing.

Per-second rates match Kie Vidu Q4 wholesale COGS, tuned for ~20–30% gross margin at Max yearly ($0.024/credit). Image-to-video and reference-to-video bill the same.

5s proof · 720p

At 720p $0.48

20 credits

Image or reference mode

12s dialogue · 1080p

At 1080p $1.43

60 credits

Reference to video

8s hero · 2K

At 720p $1.52

64 credits

Image to video

16s final · 4K

At 4K $6.47

272 credits

Max duration tier

Yearly plans lower your cost per Vidu Q4 clip

Max yearly ≈ $0.024/credit · Monthly ≈ $0.029 · Packs ≈ $0.045. Save ~17% vs monthly and ~48% vs packs at Max yearly.

See plans and credit packs →

Resolution ladder

540p · Fastest proofs
720p · Default iteration
1080p · Social finals
2K · Hero delivery
4K · Max tier

Failed jobs may still consume credits — iterate at 720p before 4K finals.

USD estimates use Max yearly $/credit; your plan may differ.

Fix common Vidu Q4 misses

Change one knob per rerun — prompt, ref count, duration, resolution, or audio toggle.

Dialogue garbled or missing words

Lines not explicit in prompt, or too many refs.

Quote dialogue in prompt; drop unrelated reference images.

Voice does not match ref

Noisy MP3 or mixed speakers in voice sample.

Upload 3–12s clean speech; one speaker; no music.

Face or wardrobe drifts

Conflicting reference images.

Use fewer, sharper refs — separate character vs location.

Text-to-video unavailable

Vidu Q4 has no T2V SKU on Kie.

Switch to Image/Reference mode or pick another model.

Pick Image or Reference mode
  1. 1. Upload only what you need

    Skip unrelated images. For dialogue, upload a clean 3–12s voice sample and write the new lines in the prompt.

  2. 2. Write timed beats + dialogue

    Address refs by order: reference image 1, reference image 2, reference audio 1. Include camera moves and sound cues.

When to use Vidu Q4 vs other models

Q4 wins on reference images + voice refs in one native-audio pass — not on length or text-only scripts.

Kling 3.0

Kling 3.0

Longer native clips, multi-shot storyboard, action-heavy motion.

Q4 wins on reference images + voice refs in one native-audio pass — not on length or text-only scripts.

Seedance 2.5 Reference Studio

Seedance 2.5

Up to 30s, @video1 motion import, largest ref bundles.

Q4 wins on reference images + voice refs in one native-audio pass — not on length or text-only scripts.

Reference to Video hub

Reference to Video

Multi-model Reference Studio with social link import.

Q4 wins on reference images + voice refs in one native-audio pass — not on length or text-only scripts.

MiniMax H3

MiniMax H3

Text-to-video + Reference Studio with @tags and 2K finals.

Vidu Q4 FAQ

What is Vidu Q4 Preview?

Image-to-video vs reference-to-video on Vidu Q4?

How do Vidu Q4 voice references work?

Does Vidu Q4 support text-to-video?

How long can a Vidu Q4 clip be?

How are Vidu Q4 credits calculated?

Vidu Q4 vs Kling for short drama?

Vidu Q4 page vs Reference to Video hub?

Is Reeldo the official Vidu website?

Explore more

Reference to VideoImage to VideoKling 3.0Seedance 2.5MiniMax H3Video ExtendPricingStudio

Reeldo is not affiliated with Shengshu Technology / Vidu. Specs follow Kie API docs.