Image References: Make the Subject Readable
Use a sharp, well-lit image where the subject fills the frame — front-facing for people, clean angles for products.
Reference to Video: Keep Your Character, Change the Scene
Reference to video on Wandura guides a new clip with a reference image or a reference video, depending on the model. Kling O1 uses a video reference and the other models use an image; generations are paid in Wandura credits (1 credit = $0.01), and the cost is shown before you create.
Before
After
The Same Character, Every Single Shot
The hardest problem in AI video is identity drift — your character turns into someone else between clips. Reference to Video anchors each generation to your uploaded image, so the face, outfit or product design stays consistent while the scene, action and camera change with your prompt. The pair below is a real result: the character reference on the left is the exact input that produced the clip on the right — same identity, dropped into a neon-lit street.
Try Reference to VideoChoose Image Reference or Video Reference
The model you pick decides which reference you upload and what it guides. Duration, resolution, audio and cost differ by model and are shown in the composer.
| Mode | You upload | Mainly guides | Prompt describes | Example models in the picker |
|---|---|---|---|---|
| Video reference | A reference video clip | Camera motion and visual style | The new subject and scene | Kling O1 (video reference) |
| Image reference | A clear image of a person, character or product | Who or what appears | The action, setting and camera | Dreamina Seedance 2.0 (ref) and other image-reference models |
Prepare Better Image and Video References
Different modes need different references.
Video References: One Clear Move
Choose a short clip with a single, steady camera move. Cuts and chaotic motion give the model mixed signals.
Only Use Assets You Can Use
Use references you own or have permission to use, and avoid recreating real people without their consent.
5 AI Models, One Subject → Video (Ref) Tool
Bring a Brand Mascot to Life
Upload one clean render of your mascot and generate an entire content series — waving hello, unboxing a product, dancing for a holiday post — with the character staying perfectly on-model in every clip. The clip here is a real generation from the single reference shown above.
Try Reference to Video
A Consistent Virtual Creator for UGC
Build a recognizable face for your channel without hiring talent. One reference portrait — like the example input shown here — becomes dozens of talking-head and lifestyle clips: same person, new scene each time, ready for TikTok and Reels in 9:16.
Try Reference to VideoProducts in Motion, Identity Intact
Logos and product designs must not warp. Anchoring the generation to your product shot keeps labels and shapes faithful while the video adds rotation, hands, environments and camera moves.
Try Reference to VideoEpisodic Stories with a Locked Cast
Kids’ stories, comics and narrative shorts fall apart the moment the hero changes face between scenes. Reuse the same reference across every scene of the script, vary only the motion prompt, and cut the clips into a coherent episode.
Try Reference to VideoPreviz and Pitch Videos with a Consistent Lead
Selling a concept is easier when the character survives from shot to shot. Generate every beat of the pitch — intro, product moment, closing scene — from one reference image and present a storyboard that already moves.
Try Reference to VideoHow It Works
- Step 01
Step 1: Upload a Reference
Add a clear image of your subject — a character, a person or a product. This is the identity the video will preserve.
- Step 02
Step 2: Describe the Motion
Write what the subject does — "walks forward, looks at camera". Pick aspect ratio, resolution and duration (4–10s).
- Step 03
Step 3: Generate the Video
The model animates your subject into the scene. Chain the result into Video Upscale, Extend or Lipsync to finish the shot. Making several shots of one character? Start with a sheet from the AI Character Generator and follow how to keep a character consistent in AI video.
How Much Does Reference to Video Cost?
Every model below runs on one Wandura credit balance — pay only for what you generate, with no separate subscription per model.
| Model | Pricing |
|---|---|
 Kling O1 (video reference) | Clone motion/style from video |
 Dreamina Seedance 2.0 (ref) | Identity-preserving |
| Kling 3.0 | Cinematic |
 Veo 3.1 | Premium · audio |
| Kling 2.5 Turbo | Draft · fast |
FAQs
It generates a new video guided by a reference: an image that carries the subject’s identity — face, outfit, product design — or a video that carries camera motion and visual style, depending on the model you pick.
It depends on the model. Kling O1 (video reference) takes a reference video; the other models in the picker use a reference image. A paid account starts on Kling O1 (video reference); a signed-out visitor or a free account starts on Kling 2.5 Turbo, the model a free first video can use. The upload slot changes with the model you select, so pick the model first.
An image reference mainly guides who or what appears — a character’s face and outfit, or a product’s look. A video reference guides how the shot moves — camera motion and visual style — while your prompt describes the new subject or scene.
Image to Video animates the exact photo you upload. Reference to Video uses your image as an identity anchor and can place the subject into new actions and framing described by your prompt. To animate the exact photo you upload instead, open Image to Video.
Reference to Video makes a new clip that keeps a character, outfit or product consistent while your prompt changes the scene, action and camera. Mimic Motion instead makes the character in your photo copy the dance, gesture or walk from a reference video.
The tool runs on Kling O1 (video reference), Dreamina Seedance 2.0 (ref), Kling 3.0, Veo 3.1, Kling 2.5 Turbo — the same lineup as Image to Video, switchable per generation.
Each model shows its own duration, resolution, audio options and credit cost in the composer. Check them after you choose the model rather than assuming one set of limits applies to all.
With an image reference, describe the action and setting — “walks through a neon-lit street at night, camera tracking alongside”. With a video reference, describe the new subject and scene; the camera motion and style come from the clip.
A clear, well-lit image where the subject fills most of the frame — front-facing for people and characters, clean angles for products. The more readable the identity in the reference, the more faithfully it carries into the video.
Dreamina Seedance 2.0 (ref) is the identity-preserving specialist in the lineup — pick it when staying on-model matters most. Kling 3.0 and Veo 3.1 trade a little identity strictness for more cinematic motion and image quality.
Yes — set the Videos count up to 4 per run. All takes use the same reference and prompt, land together in your Creations, and you can chain whichever one reads best straight into the next tool.
Yes — that is the point. Keep re-using the same reference image with different motion prompts and you get an episodic series where the character stays consistent across every clip.
Faces, outfits, small logos, text on products and props can change between generations, even with the same reference. Review these details in every clip, especially when you build a series.
Chain it forward without leaving Wandura: upscale the clip, extend it, add lipsync for dialogue, or generate a matching soundtrack with Auto Sound. Everything is saved to your Creations.
With Kling O1 (video reference) you add a reference video, plus up to 4 optional product or avatar reference images.
Premium Exclusive models (Kling 3.0, Veo 3.1) are the latest top-tier generation models. Every plan (Basic and Growth) can use them on a pay-per-use credit basis.
You can open the tool and add your files without signing in. When you press Generate, Wandura asks you to sign in or create an account. A free account gets one free generation, and here it can be one Kling 2.5 Turbo clip of up to 5 seconds (45 credits, taken from the credits the free account starts with, with a watermark); other models and longer clips use the credits of a paid plan.
Both plans do. Basic is $15/month with 1,500 credits and Growth is $49/month with 6,000 credits, billed monthly. Credit packs top up a plan and are valid for 90 days.
The credit cost depends on the model and the settings you pick, such as video length or resolution. The composer updates the price as you change settings and shows it before you create (1 credit = $0.01).
Transparent pricing
$0.01 / credit — no hidden per-model cost
70+ AI models in one studio
200+ use cases · 1000+ templates
Secure payment via Stripe
Encrypted checkout
Cancel anytime
Keep access to the end of your cycle