Skip to main content

How to Get Started with Photo Avatars

Learn how to make your photos come to life with this guide on Photo Avatar creation.

Avi avatar
Written by Avi
Updated over a week ago

A Photo Avatar lets you bring a still image to life! Simply upload a photo, and we’ll transform it into a dynamic, animated version that lip-syncs perfectly to your script. With natural expressions and realistic movements, it feels like the person in the photo is actually speaking. It’s a fun, engaging way to make your content stand out.

Photo Avatar Limits by Plan*:

Free [Mobile]

Free [Web]

Creator

Team+

Photo Avatar Slots

3

3

Unlimited

Unlimited

Looks Per Slot [Photo + Video Avatar Combined]

500

500

500

500

Training

1/month

0

1 = 20 Gen Credit

1 = 20 Gen Credit

Looks Generations

40 per calendar month

40 per calendar month

1 = 1 Gen Credit

1 = 1 Gen Credit

Add Motions or Upscale

3 per calendar month

3 per calendar month

1 = 10 Gen Credits

1 = 10 Gen Credits

Avatar IV Video

10s/video | 30s/month

10s/video | 30s/month

1 = 20 Gen Credits

1 = 20 Gen Credits


Upload Your Own Photo Avatar or Generate an AI Photo Avatar

On your HeyGen homepage, go to the Avatars tab, and choose New Avatar, where you can select to upload your existing photos or generate a photo-based virtual character with prompts and describing exactly what you're looking to create. To create your own unique photo-based character choose Design with AI option.


Tips for creating a virtual character when uploading photos

When trying to create a virtual characters with your own uploaded-photos, there are a few key elements to note to make sure you can produce a very natural and expressive photo avatar that works best with our avatar engines.

Photo Avatars work best when the character’s face is close to human proportions. Even if your character is an animal, robot, or alien, it can still work great as long as the face structure is human-like or close to human proportions.

✅ Works best (recommended)

  • Human-like face proportions

  • Eyes are clearly visible (normal-ish size and placement)

  • Mouth and lips are clearly visible

  • Face shape is easy to recognize (clear forehead / cheeks / jaw area)

  • Character looks “human-like” in structure, even if it’s a fantasy character
    (animal / alien / robot is fine as long as it still has a readable face)

⚠️ Usually doesn’t work well (not recommended)

  • Eyes are very small or hard to detect

  • Lips aren’t visible or the mouth is unclear

  • Beaks / snouts / heavy masks that hide the mouth area

  • Faces that are too stylized or exaggerated compared to real human proportions
    (can lead to weaker facial movement + less accurate lip sync)

This characters for example are lacking a nose or easy to identify lips in order to create the realistic lip sync when the character is speaking in a video.

⚠️ Another common issue: the character is too small in the picture. The character should take up a larger portion of the image in order to work properly.


Photo Avatar Prompting Best Practices

Creating effective AI image prompts is essential for generating high-quality, visually appealing, and accurate images. A well-crafted prompt ensures the AI understands your expectations and produces the desired output. This guide outlines best practices for crafting detailed and specific AI image prompts.

Structure of a Strong AI image prompt

A strong AI image generation prompt should follow four key components in this order:

  • Image type: Specify the kind of image you want (e.g., photo, product shot, illustration)

  • Main Subject: Clearly describe the central focus of the image (e.g., product, person, animal, object)

  • Background scene: Provide context by describing the setting (e.g., studio, canyon, city street)

  • Composition style: Add modifiers to describe mood, lighting, and aesthetic (e.g., natural lighting, warm tones, ultrarealistic)

You can also define the orientation (landscape, portrait, square) and style (realistic, cinematic, vintage, cyberpunk, etc.).

Example Prompts:

Basic: "A product shot of a car on a road."

Improved: "A product shot of a luxury car parked on the side of a road, next to a canyon with a flowing river, in elegant, natural lighting."

Basic: "A photo of a plant on a table."

Improved: "A photo of a succulent in a ceramic pot, placed on a wooden table in a minimalist room with soft natural lighting and warm tones."

Tips for Prioritizing Details

AI gives greater emphasis to the earlier parts of a prompt. To ensure the most important elements are prioritized:

1. Start with the image type and main subject

2. Follow with the background scene and composition style

3. Add extra modifiers at the end to refine the output

Example:

"A photo of a golden retriever playing in a sunny park, surrounded by blooming flowers, with vibrant colors and natural lighting."

Using Iterative Prompting

AI image generation is an iterative process. Start with a basic prompt and refine it gradually to achieve the desired output. Add or adjust modifiers to tweak mood, lighting, or composition, and generate multiple outputs to select the best.

Steps for iterative prompting:

  1. Start simple: "A photo of a cat on a sofa."

  2. Add more details: "A photo of a white Persian cat lounging on a plush gray sofa in a modern living room, surrounded by indoor plants, with soft natural lighting."

  3. Refine further: "A photo of a white Persian cat lounging on a plush gray sofa in a modern living room, surrounded by indoor plants, with soft natural lighting and warm tones."

Prompt Starters

Unsure where to start? Here are a few sample prompts to get you going!

Realistic Photo Sample Prompt

"A product shot of [product] on a [surface type] in a [studio setting], [composition style]."

Illustrative artwork sample prompt

"A flat illustrative artwork of [main subject] in [background], [composition style]."

Interior design sample prompt

"A photo of [object] in a [specific room], surrounded by [furniture and decor], [composition style]."


Common Photo Avatar Prompt Issues and Solutions

Below are suggestions for how to improve the most common issues our users face when generating images and avatars.

Mismatch between prompt and output

Ensure prompts are detailed and structured. Avoid vague language and include specific modifiers.

Low-quality images

Add terms like "high resolution," "sharp focus," or "ultrarealistic" to emphasize clarity.

Artifacts or distortions

Use terms like "clean rendering" or "artifact-free." Refine iteratively to reduce distortions.

Inconsistent style

Clearly state the desired style (e.g., realistic, cinematic, vintage) and avoid mixing multiple styles.

Lighting issues

Add lighting details like "natural lighting" or "warm tones" for better results.

Missing details

Include "highly detailed" or "intricate" to ensure key elements are included and accurate.

Incorrect details

Be specific about every element in your prompt and refine it iteratively if the output includes errors.

Unrealistic objects

Use terms like "realistic" or "lifelike" to guide the AI toward plausible renderings.

Bad cropping

Include framing instructions such as "wide shot," "mid-range" or "close-up" to control how the subject is positioned in the frame.

Unnatural poses or expressions

Specify "natural pose" or "authentic emotion" to improve realism.


Creating Videos with Your Photo Avatar

Next, head over to the Create Video tab and select Avatar Video. Choose your desired video orientation, then go to Avatars and select Photo Avatar. Pick your preferred photo avatar and add it to your work board.

From here, head to the Script tab to write what you'd like your photo avatar to say. For more details on scripts, check out this Script article. You can also choose a 'Voice' for your photo avatar in the script section - more details on Voices can be found in this Voice article.

Finally, you can add visual elements or text to make your video even more dynamic by navigating to the Text and Element tabs.


Adding Motion to Photo Avatars

You can now enhance your photo avatar with different motion engines, directly from our studio! To add motion, click choose your photo avatar inside studio, and click on the Avatar Tab. under Motion Engine, click on the small arrow to open up the list of available engines ( users in different regions might see more or less engines than others, but Unlimited and AvatarIV should always be available).

You can also click on the Advance Settings button to add prompt based custom motion that might meet your video needs best.


Please mind all engines other than Unlimited use up your GenCredits, so please review our GenCredits article for more.

  • Unlimited engine uses no GenCredits and no limit on length of scenes.

  • AvatarIV uses 20 GenCredits for every 1 minute of video, and is limited to 180 seconds scenes. You can make a longer video than 180 seconds, you just need to split each script-box to not exceed 180 seconds.

  • Kling, Runway, Hailuo, Seedance - any video created with these engines would use up 10 GenCredits, regardless of the length of the video.


🛎️ Some users may still have the option to add motion to the photo avatar outside of Studio ( see clip below). This option is currently being depreciated and our team cannot provide access to this option for users who don't have it on their account. 🛎️

Motion Photo Avatar Prompting Best Practices

Use clear and direct descriptions

Avoid abstract or conceptual phrasing and instead use concrete descriptions.

❌ A person attempting to bypass security on a computer
✅ A person rapidly typing lines of code on a computer screen.

Avoid conversational or command-based prompts

Generative video models interpret clear, visual descriptions better than conversational phrases or commands.

❌ can you please make me a video about a cat walking away from her kittens?
✅ a cat walking away from her kittens

Similarly, avoid direct commands:

❌ add snow to the image
✅ snow covering the field from out of frame

Use Positive Phrasing

Negative prompts (what shouldn’t happen) are not well-supported and may result in unintended effects.

❌ The scene remains unchanged. No motion. An empty sky.
✅ Fixed camera. The shot stays steady. A bright, cloudless sky.

Prompt Keywords

Keywords can help achieve specific styles or effects in your output. Ensure your keywords align with the overall context of your prompt.

For instance, keywords about skin texture are more relevant for close-up shots, while a wide-angle scene might benefit from environmental details.

Experiment with the following categories to refine your prompts:

  1. Mood: Stunning, radiant, delicate, striking, glamorous

  2. Lighting: Warm lighting, natural lighting, cold lighting, dark aesthetic

  3. Position: Centered, off to the side, on the left, on the right

  4. Detail: Highly detailed, intricate, ultra-realistic

  5. Angle/Viewpoint: Close-up, headshot, 3/4 shot, wide shot, low angle shot, high angle shot

By incorporating clear descriptions, positive phrasing, and cohesive keywords, you can craft prompts that consistently deliver high-quality results.


Generating Different Looks

Interested in a Video Tutorial? Check out our Generating Looks tutorial below.

Once you have your base photo avatar, you can continue generating different looks for your avatar in three different ways-

  • Uploading more photos as reference for our engine to reference when generating new looks

  • Prompting and describing which looks you want to create (Pro tip- it's best to finetune your prompt with Gemini, ChatGBT, or another LLM)

  • Choosing "inspiration" photos from our library.

All of these three options to generate more photo looks are available under the avatar main tab. When you choose your photo avatar in our main avatar tab, these three options will appear in the bottom of the screen as in the clip attached below.


Frequently Asked Questions (FAQ)

How many photo avatars can I create?

The amount of photo avatars a user can create depends on their subscription. Free users can create up to 3 unique photo avatars, while Creator, Team and Enterprise users have Unlimited photo avatar slots.

Why isn't my photo avatar speaking after I submit my video?

It's possible you uploaded your photo as an 'asset' instead of as a photo avatar. Ensure you are following the instructions above for proper photo avatar creation. If you followed the instructions above and are still not seeing the proper output, please reach out to [email protected] for assistance.

Why did my photo avatar fail / not succeed?

There are many reasons your photo avatar may have failed. See this article for a list of common failure reasons. If you still have questions, please reach out to [email protected] for a manual review.


*Last Updated: Dec 8, 2025

Did this answer your question?