HeyGen Reference-to-Video: Photo to Video

HeyGen reference-to-video is a mode of HeyGen's video model, HeyGen Video 1, in which you supply your own images (and optionally video clips and audio) as references and describe the scene in text, and the model generates a 5 to 15 second clip that uses them, with dialogue and sound effects generated in the same pass. If you only want to animate one photo you already have, a browser image-to-video generator is the shorter route: GenModelHub's image to video page animates one uploaded photo into a short MP4 with no avatar or reference set-up. GenModelHub is not affiliated with HeyGen. Updated October 26, 2026.

How to make a video from one reference photo with GenModelHub (3 steps)

This is the short route when your goal is "my photo, but moving". It runs in a phone or desktop browser.

1. Pick the photo. Open image to video and upload a PNG, JPG or WebP up to 10 MB.
2. Describe the motion. One sentence is enough: who or what moves and how, for example "she turns to the camera and smiles, slow push-in". You can also leave it empty and let the model choose natural movement.
3. Generate and download. A clip usually takes about a minute. Download the MP4 and post it.

Which photo works best

• One clear subject. One face or one object that fills a good part of the frame.
• Sharp and well lit. A daylight photo beats a dark or blurry one; the model can only animate what it can see.
• No heavy filters or sunglasses on a face you want to keep recognisable.
• Pets and objects work as well as people. Photos of minors, nudity and ID documents are not accepted.
• Good: a sharp, front-facing portrait with a plain background. Weak: a group photo where each face is tiny.

What you get, and what it costs

The image-to-video page runs MiniMax H3 and returns a 5 or 10 second 768p MP4 without sound. It costs 2 credits per second, so a 5-second clip is 10 credits. GenModelHub sells credits one time, with no subscription: $8.99 buys 30 credits and $24.99 buys 90. Failed generations are refunded automatically. If the motion is not what you wanted, change one phrase in the description and run it again.

What HeyGen reference-to-video is, in facts

• The model: HeyGen presents HeyGen Video 1 as its own video model, which "generates the picture, dialogue and sound effects of a scene in one pass" (HeyGen). HeyGen's page does not state a release date.
• Inputs: text, images or reference clips. In the HeyGen app you can "bring your own images and references" and build scenes around your avatar; the API accepts text, image or reference to video.
• Reference limits (via ComfyUI's partner node): up to 9 reference images, 3 reference videos and 3 reference audios, 12 references in total; audio needs at least one reference image or video; references are cited in the prompt as @Image1, @Video1, @Audio1 (ComfyUI docs). These are limits of that integration; HeyGen's own page gives no reference count.
• Length and size: 5 to 15 seconds per scene, 480p to 2K; 2K needs a 16:9 or 9:16 aspect ratio when references are attached (ComfyUI docs).
• Prices: HeyGen lists the API at $0.01 per second as a "launch price through October" (no year stated). A free HeyGen account gets 2 clips a day, 5 seconds each, with sound; longer clips, your own images and 2K need the HeyGen app.
• Ranking claim: HeyGen calls it "the #1 video model" on Artificial Analysis' video leaderboard. This is HeyGen's own statement; we did not check the leaderboard.
• Facts above were read on October 26, 2026 and may change; prices in particular are launch prices.

How to use reference-to-video on HeyGen

1. Prepare references. Pick clear images (a face, an object, a location). Add a short video or clean audio only if you need that motion or voice.
2. Write a shot list. HeyGen's guidance is to cover subject and action, camera, look and sound, and to state exclusions such as no music, logos or lettering.
3. Choose style and format. The playground offers Claymation, Live action, Anime and Screen print, in 16:9, 9:16 or 1:1.
4. Generate in the playground or the app. The free playground makes 5-second clips; the HeyGen app is needed for your own images, longer clips and 2K.
5. Using the API: call text, image or reference to video with seeds and webhooks.

The cost of that route: a HeyGen account, a prepared reference set and prompt, and (outside the free tier) paid use. The benefit is control: several references, sound with speech, up to 15 seconds and 2K.

Comparison

	HeyGen reference-to-video	GenModelHub image to video
Input	Up to 9 images, 3 videos, 3 audio (ComfyUI node) plus prompt	One photo plus optional motion sentence
Length	5–15 s	5 or 10 s
Resolution	480p–2K	768p
Sound	Dialogue and effects generated	No sound (add music in TikTok or your editor)
Account	HeyGen account for the free tier	Pay per clip, no subscription
Cost	Free: 2 five-second clips a day; API $0.01 per second launch price	2 credits per second; $8.99 buys 30 credits, $24.99 buys 90
Best for	Scenes with several characters or looks, speech	Quickly animating one photo

Pick HeyGen if you need several references, speech or 2K. Pick the browser generator if you have one photo and want a short silent clip quickly.

If the result doesn't look right

Problem	Fix
The face drifts or changes	Use a sharper, larger, front-facing photo and describe less motion
Motion is jittery	Describe one action and one camera move
Wrong look or style	Name the style in the prompt (live action, anime, claymation)
Need speech or sound	Use a tool with generated audio, or add a track when you post
Clip too short	Make several clips from the same photo and join them in an editor

FAQ

What is HeyGen reference-to-video? It is a mode of HeyGen Video 1 that generates a 5 to 15 second clip from your own reference images, videos and audio plus a text prompt, with dialogue and sound effects created in the same pass.

How do I make a video from a reference photo? On HeyGen, add the photo as a reference in the app or API and describe the scene. On GenModelHub, open image to video, upload one photo, describe the motion and download the MP4 after about a minute.

Is HeyGen reference-to-video free? HeyGen says a free account gets 2 clips a day, 5 seconds each, with sound; your own images, longer clips and 2K need the HeyGen app. The API is listed at $0.01 per second as a launch price through October.

How many reference images can I use? HeyGen's page gives no count. The ComfyUI partner node documents up to 9 images, 3 videos and 3 audio files, 12 in total.

Is GenModelHub the official HeyGen site? No. GenModelHub is an independent service, not affiliated with HeyGen. HeyGen and HeyGen Video are names of HeyGen products, used here only to describe them.

More: compare every model's cost and make a trend video from two photos.

If what you need is a product video ad rather than an avatar video, see the Arcads alternative and Creatify alternative pages.

Related guides: ai ad maker, ai video ad generator and ai commercial maker.