What is reference-to-video?

With reference to video, you upload one or more images of a subject (a character, a product, an outfit, a place, or a visual style) and write a prompt for a brand-new scene. The model keeps the subject from your references and builds the motion, background, and camera work around it.

Reference-to-video vs image-to-video:

• Image-to-video animates one specific photo. That photo becomes the first frame, so you stay in that scene.
• Reference-to-video uses your images as a guide, not a first frame. The subject can appear in a whole new setting, pose, or camera angle, which is what you want for series, ad campaigns, and storyboards.

A quick rule: if the job is "make this picture move," use image-to-video. If it's "put this character or product somewhere new," use reference-to-video.

Which models support it on Veemo

When we checked the generator (October 2026), these models had a reference-to-video mode. Each one accepts different inputs:

If you only have still images, start with MiniMax H3, Seedance 2.5, or Veo 3.1. The Seedance 2.0 models and Kling O3 expect a video input as well.

Step-by-step workflow

1. Make or pick clean reference images. Use sharp, well-lit images where the subject fills a good part of the frame. If you don't have any yet, generate them with Nano Banana Pro. It supports multiple reference images and character consistency for up to 5 people, and outputs at 1K, 2K, or 4K. A front view, side view, and full-body shot on a plain background work well.

2. Choose a model. Pick based on your inputs (images only, or video too), clip length, and budget.

3. Upload your references. Add more images only when each one adds something, such as a side profile or the outfit.

4. Write the prompt. Say what each reference is, then describe the scene, action, camera, and light. The images already carry the identity, so keep the subject description short.

5. Set the format. Choose aspect ratio (9:16 for Reels and TikTok, 16:9 for YouTube), length, and resolution. The cost is shown before you generate.

6. Test short, then scale. Run a short, low-resolution test. If the identity holds, reuse the references and change only the scene prompt.

Copyable prompts

Change the bracketed parts to match your project.

1. Character consistency: new location

Use the woman from reference image 1 as the main character. Keep her face, hair, and red coat exactly the same. She walks through a busy night market in Tokyo, glancing at food stalls, warm lantern light, handheld camera following from behind, then she turns toward the camera and smiles.

2. Character consistency: series episode

Same character as the references: [short description]. Morning scene in a small sunlit kitchen. He pours coffee, looks out the window, and checks his phone. Medium shot, soft natural light, calm cozy mood, realistic style.

3. Product ad: hero shot

The product is the bottle from the reference images. Keep the label, cap, and color exactly as shown. The bottle stands on wet black stone, water droplets slide down the glass, slow 180-degree orbit around it, dramatic rim lighting, premium commercial look.

4. Product ad: lifestyle

A young woman holds the sneaker from reference image 1 and laces it up on a city rooftop at sunset, then jogs toward the camera. Keep the shoe's colors and logo unchanged. Vertical 9:16, energetic social ad style.

5. Style transfer

Match the visual style of reference image 2: [hand-painted watercolor / 1980s anime / claymation]. The character from reference image 1 rides a bicycle along a coastal road, wind in her hair, wide tracking shot. Keep the character's design consistent with reference image 1.

6. Multi-character scene

Reference image 1 is the man with the gray beard, reference image 2 is the girl with the yellow raincoat. They sit together on a park bench in light rain and share an umbrella. He tells a story, she laughs. Keep both characters looking exactly like their references. Medium two-shot, gentle camera drift.

7. Character plus product

The character from reference image 1 unboxes the headphones from reference image 2 at a desk, puts them on, and nods to the music. Keep both the character and the product design consistent. Bright home studio, YouTube review style.

Tips and common mistakes

• Messy references. Busy backgrounds, small faces, heavy filters, and sunglasses all weaken the result.
• Fighting the references in the prompt. If the reference shows a red coat and the prompt says "blue jacket," expect drift. Describe the scene, not a new look.
• Not labeling your images. With more than one reference, say "reference image 1 is..." so the model knows which subject is which.
• Extreme angles and lighting. Looking straight up, full profiles, and deep shadows are where details slip most.
• Long clips before tests. Check identity on a short clip before paying for 15 or 30 seconds.
• Changing everything at once. For a series, keep the references and style wording fixed and change only the action and location.

How much does it cost?

Veemo uses credits, and the generator shows the exact cost before you generate. Reference-to-video prices when we checked:

• MiniMax H3: 4 credits per second at 768P, 7 per second at 2K. A 4-second 768P clip costs 16 credits.
• Seedance 2.5: 22 credits per second at 480p, 47 per second at 720p (less when you add a reference video).
• Veo 3.1: 40 credits per generation at 720p, 50 at 1080p.
• Wan 2.6: 80 credits for 5 s at 720p, 150 for 5 s at 1080p.

Free credits: new verified accounts get 16 credits, valid for 30 days, with no payment details required. That covers a 4-second MiniMax H3 clip at 768P, which is plenty for a first test. Downloading a result requires a credit pack or plan.

Paid options: subscriptions start at $29.90/month (598 credits) and annual billing is 50% off. One-time credit packs start at $9.90 for 152 credits, and pack credits never expire.

FAQ

How many reference images can I upload? It depends on the model: up to 3 on Veo 3.1 and Wan 2.6, up to 9 on MiniMax H3, and more on Seedance 2.5. The upload box shows the limit.

Can I use the same character in many different videos? Yes, that's what it's for. Keep the same references and change the scene prompt for each clip.

Do I need a video as a reference? No. MiniMax H3, Seedance 2.5, and Veo 3.1 work with images only. The Seedance 2.0 models and Kling O3 need a video input in this mode.

Can I use the videos commercially? Veemo's reference-to-video page says the content you generate comes with full commercial rights and needs no attribution.

Ready to try it? Open Veemo reference-to-video, upload your best reference, and run a short test.

This guide covers a prompt structure that works, eight example prompts you can copy, and how to use the result as a keyframe for Seedance 2.5 image-to-video.

Nano Banana vs Nano Banana Pro

There are a few versions of the model. The original Nano Banana is fast and cheap: good for mood boards, sticker packs and lots of variations. Nano Banana Pro is the one to use when you need precise text rendering, multiple references, character consistency or 4K output. Everything below works on both, but the editing and typography examples benefit most from Pro.

The prompt structure that works

Write in full sentences, roughly in this order:

1. Subject: who or what, with the specific details that matter.
2. Setting and composition: where it is, framing, what's in the foreground and background.
3. Camera and lens: shot size, angle, depth of field.
4. Lighting and color: time of day, light source, grade.
5. Style and finish: photographic, illustrated, 3D, etc.
6. Text (if any): put the exact words in quotes and say where they go.
7. Constraints: what must stay the same or must not appear.

For edits, flip it around: say what to keep first, then the single change you want. Small, specific edits work much better than "make it better".

8 example prompts you can copy

1. Product hero shot (text-to-image)

A matte black ceramic coffee mug on a light oak table, morning sunlight coming from the left through a window, soft shadows, a thin wisp of steam rising from the cup. 50mm lens, eye level, shallow depth of field with the background softly blurred. Clean, warm commercial photography, neutral color grade. No text, no logos.

2. Poster with readable typography

A minimalist travel poster for Kyoto in autumn: a single red maple branch in the upper right corner, a pagoda silhouette at the bottom, cream paper texture background. The headline "KYOTO" in bold geometric sans-serif at the top center, and the smaller line "Autumn 2026" below it. Flat illustration style, limited palette of red, charcoal and cream.

Quoting the exact text and saying where it goes is the main thing that gets clean lettering out of Pro.

3. Consistent character, new scene

Upload a reference image of your character, then:

Keep the same woman from the reference image: same face, hairstyle and green raincoat. Place her on a rainy Tokyo street at night, holding a clear umbrella, neon signs reflecting on the wet pavement. Medium shot, cinematic lighting, slight teal-and-orange grade.

4. Single-change edit (image-to-image)

Keep everything in this photo exactly the same, but change the sky to a soft pink sunset with scattered clouds, and adjust the lighting on the building to match the warm sunset light.

5. Multi-image fusion

Upload a product photo, a background photo and a style reference:

Place the sneaker from image 1 on the concrete steps from image 2, lit to match the scene. Use the color palette and high-contrast mood of image 3. Low angle, the sneaker sharp in the foreground, background slightly out of focus.

6. Infographic

A clean vertical infographic titled "How Coffee Is Made" with four numbered steps: "1. Harvest", "2. Dry", "3. Roast", "4. Brew". Each step has a simple flat icon and a one-line label. White background, brown and orange accent colors, modern sans-serif type, generous spacing.

7. Character sheet for animation

A character turnaround sheet of a small friendly robot with a round body, one large blue eye and stubby arms: front view, side view, back view and three-quarter view in a row, plain light gray background, consistent proportions and colors in every view, soft 3D render style.

8. Cinematic keyframe (made for video)

Wide cinematic shot of a lone hiker in a red jacket standing on a mountain ridge at sunrise, low clouds rolling through the valley below, golden light on the peaks. The hiker is small in frame on the left third, facing right toward open space. 21:9 composition, natural colors, high detail, no text.

The next section explains why this one is written the way it is.

Turning your image into a Seedance 2.5 video

A strong still makes a strong first frame. Here's the workflow I use:

1. Compose for motion. Leave room in the direction the subject will move or the camera will travel. In prompt #8 the hiker faces open space on the right, so a slow pan or push-in has somewhere to go. Avoid heavy text in the keyframe. Video models can warp lettering once it starts moving.

2. Match the aspect ratio. Generate the image at the ratio you want for the video. Nano Banana Pro supports 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3 and 21:9. Seedance 2.5 supports 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 and adaptive. Seedance renders at up to 720p, so a 1K or 2K image is plenty for a keyframe. Save 4K for print or stills.

3. Write a motion prompt, not a picture prompt. The image already defines the look. Describe what changes:

Slow push-in toward the hiker as the clouds drift through the valley, the sun rises slightly and the light warms on the peaks, the jacket moves gently in the wind. Soft wind ambience. Steady, cinematic camera.

4. Choose length and resolution. Seedance 2.5 lets you pick any whole number of seconds from 4 to 30 at 480p or 720p. Draft at 480p for 4–6 seconds, then render the version you like at 720p.

5. Use First & Last Frames for transformations. Want the mug from prompt #1 to go from empty to full, or a room to go from day to night? Make both states with Nano Banana Pro (use prompt #4's "keep everything the same" pattern for the second one), then upload them as first and last frame in Seedance 2.5. The model fills in the motion between them.

6. Chain shots. Turn on "Return Last Frame" in Seedance 2.5 and you can reuse the final frame as the starting image for the next clip.

If you want 2K video or native stereo sound on a short 4–15 second shot, the same keyframe also works with MiniMax H3.

Cost and getting started

On Veemo, a Nano Banana Pro image costs 30 credits at 1K or 2K and 40 credits at 4K. Seedance 2.5 video is billed per second: 22 credits/s at 480p and 47 credits/s at 720p. New accounts get 16 free credits after verification (valid 30 days), enough to try a short 4-second MiniMax H3 clip. Subscriptions start at \(29.90/month (598 credits), or you can buy one-time credit packs that never expire, starting at \)9.90.

Common mistakes

• Keyword soup. "Mug, 8k, ultra, masterpiece" does less than one clear sentence.
• Too many edits at once. Change one thing per pass, then iterate.
• Unquoted text. If you want specific words in the image, put them in quotes.
• Busy keyframes. Clutter in the still becomes flicker in the video.

FAQ

Can I use Nano Banana Pro images commercially? Yes, they can be used for marketing, advertising and products. Generated images include a SynthID watermark that marks them as AI-generated.

How many reference images can I use? Nano Banana Pro is built for multi-image fusion (Google lists up to 14 references). The Veemo generator shows the upload limit for the mode you're in.

What resolution should I use for a video keyframe? 1K or 2K is enough, because Seedance 2.5 outputs at up to 720p.

Do I need to describe the image again in the video prompt? No. Describe the motion, camera and sound. The image already covers the look.

Is there a free way to try this? Yes. New Veemo accounts get 16 free credits after verification.

This is a practical comparison, not a benchmark. I'm not going to give you made-up scores. I'll go through the published specs, what each model is built for, and how I'd choose between them on a real project. Where something depends on your prompt or is not publicly documented, I'll say so.

The specs at a glance

The table uses the settings exposed on Veemo, which runs both models, plus Kuaishou's public documentation for Kling.

*Which resolution you get depends on the platform and mode. Kling's 4K mode is in Kuaishou's own API. Check the generator you use before you plan a 4K delivery.

Quality and look

I won't pick a winner on visual quality. It depends too much on the subject and the prompt, and anyone who says one model is "30% better" ran a test you can't reproduce. A few qualitative notes are still useful:

• Kling 3.0 is pitched for cinematic realism and steady characters. Kuaishou's launch notes focus on photorealism, keeping elements consistent across frames, and multi-shot storytelling. If your deliverable is 1080p, Kling has the higher native resolution of the two on Veemo.
• Seedance 2.5 tops out at 720p on Veemo. In exchange you get much more control over how the shot is built. You can feed it image, video and audio references, edit an existing clip, or set both the first and last frame and let the model animate the motion between them.

My rule of thumb: if the final file must be crisp 1080p straight out of the model, start with Kling 3.0. If you're assembling a sequence and need the shots to match a reference, Seedance 2.5's workflow tools usually matter more than the extra pixels.

Motion and camera

Both models respond to cinematography language. Kling's own guidance says it handles terms like "tracking shot", "close-up" and "golden hour lighting" well. In my experience the same vocabulary works for Seedance too.

The structural difference is First & Last Frames in Seedance 2.5. You upload a starting image and an ending image and describe the motion. That's the most reliable way I know to get a specific camera move or transformation, because the model knows where the shot has to land. Seedance can also return the last frame of a clip, so you can feed it into the next generation and chain shots together.

Kling 3.0's motion strengths are of a different kind: multi-shot storytelling within one generation (per Kuaishou) and lip-synced dialogue, which matters a lot for talking-head and character scenes.

Prompt-following

Each model rewards a different style of prompt:

• Kling 3.0 likes a clear, cinematic description: subject, action, camera move, lighting, plus any dialogue you want spoken. For image-to-video, describe the motion and audio rather than re-describing the picture.
• Seedance 2.5 gives the best results when you combine a prompt with references. Use images to pin down the subject, a video for motion or style, audio for rhythm. Then keep the text prompt focused on what should happen.

One caveat from Veemo's Seedance 2.5 page is worth repeating: Extend is prompt-guided and doesn't guarantee a frame-exact seamless continuation. If you need a perfect match at the cut, use First & Last Frames or the last-frame return instead.

Length and resolution

This is the clearest difference between the two:

• Need a 20–30 second continuous clip? Only Seedance 2.5 does that in one generation (up to 30 s).
• Need 1080p? Kling 3.0.
• Need 21:9 cinematic or 4:3/3:4 social formats? Seedance 2.5 supports them. On Veemo, Kling 3.0 offers 16:9, 9:16 and 1:1.

Speed

I don't have controlled timing numbers, and render times change with server load, length and resolution, so I won't quote seconds. What's documented is that Kling 3.0 has a Standard mode meant for faster iteration and a Pro mode for final quality. For Seedance 2.5, the practical way to speed up iteration is to draft at 480p and short durations, then re-run the winner at 720p.

Cost (credits on Veemo)

Both models bill per second on Veemo, so it's easy to estimate a shot before you render it:

So per second at high quality, Kling 3.0 is cheaper. Seedance 2.5 at 480p costs the same as Kling Standard and works well for drafts. The generator always shows the total before you click Generate, so check it there. These are the rates as I write this.

So, when should you pick which?

Pick Seedance 2.5 when:

• You need clips longer than 15 seconds, or unusual aspect ratios like 21:9.
• You want to control the shot with reference images, video or audio.
• You're editing or extending existing footage.
• You need a defined start and end (logo reveals, product transformations, before/after shots).

Pick Kling 3.0 when:

• You need 1080p straight from the model.
• The scene has spoken dialogue and you need lip-sync, including in Chinese, Japanese, Korean or Spanish.
• You want cheap, fast drafts in Standard mode and a polished Pro render at the end.

Consider a third option for short, heavily directed shots: MiniMax H3 does 4–15 second clips at up to 2K with native stereo audio, and takes text, image, video and audio as references.

Trying them side by side

The simplest way to decide is to run the same prompt through both. On Veemo, the Seedance 2.5 and Kling 3.0 generators are in the same interface, so you don't need separate accounts. New accounts get 16 free credits after verification (valid for 30 days). That's enough for a 4-second MiniMax H3 clip at 768P, which gives you a feel for the workflow. Paid subscriptions start at \(29.90/month for 598 credits, and one-time credit packs that never expire start at \)9.90. If a generation fails, Veemo returns the credits.

Tip: for image-to-video tests, generate one keyframe with Nano Banana Pro first and send the same image to both models. Then the only thing that changes is the video model.

A good test prompt for both models:

A barista pours latte art in a sunlit café, slow push-in from mid shot to close-up on the cup, warm morning light, shallow depth of field, soft café ambience and the hiss of a steam wand.

Watch three things: whether the camera move matches the prompt, whether the hands and the liquid look physically believable, and how well the audio matches the action.

FAQ

Is Seedance 2.5 better than Kling 3.0? Neither one wins outright. Seedance 2.5 gives you longer clips (up to 30 s), more aspect ratios and more reference and editing workflows. Kling 3.0 gives you 1080p, lip-synced multilingual audio and a lower per-second cost at its top quality setting.

What's the maximum length for each? Seedance 2.5: 4–30 seconds. Kling 3.0: 3–15 seconds.

Can I try them for free? Yes. New Veemo accounts get 16 free credits after verification, enough for a short MiniMax H3 video. Longer Seedance or Kling renders need more credits from a plan or a credit pack.

Can I animate my own image? Yes, with both. Seedance 2.5 also supports First & Last Frames if you want to set both ends of the shot.

Who runs Veemo? Veemo is operated by AIVIDO LTD, based in London, UK.