Question 1

What exactly does Reference to Video do?

Accepted Answer

Reference to Video lets you upload one photo of a character, then generate multiple videos where that same character appears in completely different scenes. The AI does not animate your photo — it creates entirely new videos featuring your character in new environments. Think of it as giving the AI a casting photo: every video it generates will star that person, placed in whatever scene you describe in your prompt.

Question 2

What is the difference between Reference to Video and Image to Video?

Accepted Answer

Image to Video animates your photo directly — your uploaded image becomes the first frame, and the AI adds motion to that exact scene. Reference to Video is different: it uses your photo only as a character template and generates a completely new scene featuring that character. Use Image to Video when you want to bring a specific photo to life. Use Reference to Video when you want your character transported into new scenarios, environments, or time periods entirely.

Question 3

What kind of photo works best as a reference?

Accepted Answer

Clear, well-lit photos with a single subject produce the best results. Front-facing portraits or full-body shots with a simple, uncluttered background work particularly well. Avoid group photos, heavy filters, motion blur, or images where the subject's face is partially obscured. High resolution — at least 720p — gives the model more detail to work with and significantly improves consistency across multiple generations from the same reference.

Question 4

How consistent will the character look across videos?

Accepted Answer

Veo 3.1 maintains strong visual consistency — the character's facial features, hair color, and general appearance stay recognizable across all generated videos. Clothing and fine details may vary slightly depending on the scene context, since the AI adapts the character naturally to each environment. For best results, describe the character's specific outfit explicitly in your prompt and use the same reference photo for every video in a series.

Question 5

Can I use illustrations or cartoon characters as a reference?

Accepted Answer

Yes. Veo 3.1 works with photorealistic portraits, illustrated characters, game character renders, and stylized artwork. The model extracts the core visual identity from your reference regardless of the art style. Illustrated characters with strong, distinct features — unique hair, distinctive clothing, a clear facial structure — tend to maintain consistency particularly well across multiple video generations.

Question 6

How many videos can I generate from one reference photo?

Accepted Answer

There is no limit on the number of videos you can generate from a single reference photo. Each generation uses credits from your plan, but you can reuse the same reference image as many times as you want. Many creators upload one character photo and produce an entire series — placing the same character in dozens of different environments, time periods, and situations to build a cohesive, recognizable content library.

Question 7

Can I use generated videos commercially?

Accepted Answer

Yes — videos generated on any paid plan can be used commercially, including for ads, social media marketing, client deliverables, and branded content. Free-tier generations are for personal, non-commercial use only. Always ensure the reference photo you upload is an image you have the right to use commercially, and review Google's Veo 3.1 usage policies for any additional platform-specific restrictions.

Question 8

Is Reference to Video free to try?

Accepted Answer

Yes. Every new account receives 25 free credits on sign-up — no credit card required. You can use those credits to generate your first reference-to-video outputs and test the character consistency before choosing a plan. Veo 3.1 uses more credits per generation than lighter models due to its higher quality output. Paid plans start at $9.90 per month and include a significantly larger credit allocation for regular use.

One photo. Any scene.

一张照片，无限场景，始终同一角色

同一角色，不同世界——全部来自一张图片

如何生成角色一致的视频

上传一张参考图片

描述一个新场景

生成并重复

功能亮点

支持的 AI 模型

参考图转视频适合哪些人？

内容创作者

营销人员和品牌

游戏和 IP 开发者

教育者和故事创作者

虚拟网红

独立电影人

常见问题