How to Master Hailuo AI Subject Reference Model in 2025: Step-by-Step Tutorial
Hailuo AI’s Subject Reference Model enables the recognition of facial features from photos uploaded by users. It can then generate a character with corresponding facial traits and create video scenes featuring that character based on text prompts.
Subject Reference Model Tutorial
To generate videos using Hailuo AI’s Subject Reference Model, users must provide two key inputs:
- Subject's Facial Photo
- Prompt for Video Generation
- The first step is to click "Try Now" to enter the creation page.
- The second step is to select the Subject Reference Model, then click "Add Reference Character," upload the image, enter the prompt, click "Create", and after a brief wait, your creation will be ready.
Subject's Facial Photo
The Subject Reference Model requires a clear, recognizable photo of the subject’s face to serve as a reference for generating the character's face in the video. Currently, the model only supports facial recognition for human faces.
Hailuo AI is capable of accurately identifying facial features across various genders, ages, and ethnicities. To achieve optimal video generation results, ensure that the uploaded image meets the necessary specifications, including proper angles and lighting conditions.
Specifications
Face Count: The Subject Reference Model only supports uploading single-person photos for facial recognition.Image Size: Resolution should be no lower than 120x120, and the image file should not exceed 20MB.
Other Requirements: The face must not be obstructed, and the photo should not be out of focus or contain excessive filters or special effects.
Standard Facial Photo：
Examples of photos not suitable for reference：
Facial obstruction
Multiple faces
Heavy filter effects
Excessive blurriness
Subject angle
The subject angle in the uploaded photo should be either directly facing the camera or slightly turned to the side.- Directly facing the camera: The subject's face should be directly facing the camera, with clear facial features visible in the frame, providing the most complete facial information.
- Slightly turned to the side: The subject's facial features and facial contours should still be fully visible in the photo, avoiding profiles or angles where the facial features are not within the frame.
Examples of quality camera angles：
Facial Photo 1
Generated Video 1
Prompt 25
On an overcast day, in an ancient cobbled alleyway, the model is dressed in a brown corduroy jacket paired with beige trousers and ankle boots, topped with a vintage beret. The shot starts from over the model's shoulder, following his steps as it captures his swaying figure. Then, the camera moves slightly sideways to the front, showcasing his natural gesture of adjusting the beret with a smile. Next, the shot slightly tilts down, capturing the model's graceful stance as he leans against the wall at a corner. The video concludes with an upward shot, showing the model smiling at the camera. The lighting and colors are natural, giving the footage a cinematic quality.
Facial Photo 2
Generated Video 2
Prompt 2
On an overcast day, in an ancient cobbled alleyway, the model is dressed in a brown corduroy jacket paired with beige trousers and ankle boots, topped with a vintage beret. The shot starts from over the model's shoulder, following his steps as it captures his swaying figure. Then, the camera moves slightly sideways to the front, showcasing his natural gesture of adjusting the beret with a smile. Next, the shot slightly tilts down, capturing the model's graceful stance as he leans against the wall at a corner. The video concludes with an upward shot, showing the model smiling at the camera. The lighting and colors are natural, giving the footage a cinematic quality.
Facial Photo 3
Generated Video 3
Prompt 3
In a softly lit modern living space, the model wears a cozy off-white knit sweater paired with high-waisted jeans and minimalist white sneakers, exuding a comfortable and casual vibe. Behind her is a large glass window letting in the sunlight, which spills onto the wooden floor. The model sits leisurely on the sofa, flipping through a book, with a cup of hot coffee beside her. The camera slowly zooms in, capturing the moment she gently lifts her head and gazes out the window, her expression calm and gentle. Then, the camera circles around to her side, showing her adjusting her hair.
Facial Photo 4
Generated Video 4
Prompt 4
In a softly lit modern living space, the model wears a cozy off-white knit sweater paired with high-waisted jeans and minimalist white sneakers, exuding a comfortable and casual vibe. Behind her is a large glass window letting in the sunlight, which spills onto the wooden floor. The model sits leisurely on the sofa, flipping through a book, with a cup of hot coffee beside her. The camera slowly zooms in, capturing the moment she gently lifts her head and gazes out the window, her expression calm and gentle. Then, the camera circles around to her side, showing her adjusting her hair.
Examples of subject angles not suitable for reference：
Facial features are not fully visible in the frame
Profile view with insufficient facial features
Facial features are not fully visible in the frame
Profile view with insufficient facial features
Facial lighting
To achieve the most accurate video representation of the face from the uploaded photo, the subject’s face should avoid large areas of dark shadows or localized overexposure. The lighting should be natural, with some subtle shadow details on the face for better depth.Examples of standard facial lighting：
Facial Photo 5
Generated Video 5
Prompt 5
A cowboy rides a horse through a snowy landscape, with a rifle strapped to his back. He wears a wide-brimmed hat, and his rugged, striking facial features stand out prominently. In the distance behind him is a tall spruce forest, beyond which rise the breathtaking, snow-covered mountains. The ground is blanketed in pristine white snow as the cowboy slowly rides through the cold, open wilderness. His gaze is fixed ahead, exuding a sense of determination and bravery.
Facial Photo 6
Generated Video 6
Prompt 6
A model, dressed in a vintage apron with a ribbon tied around her neck and high-heeled ankle boots, stands behind the dimly lit wooden bar of a tavern. She holds a glass filled with whiskey in one hand, while the background features a wooden shelf lined with bottles and oil lamps. Her curly hair cascades over her shoulders, and a mischievous smile plays on her face as the camera moves dramatically.
Examples of unsuitable facial lighting for reference：
Shadows obscure the facial features
Complexed facial shadows
Excessive or unnatural facial lighting colors
Overexposure
Prompt for Video Generation
The generation logic of the Subject Reference Model is similar to that of text-to-video generation. For more details, please refer to the Text-to-Video Prompt tutorial.
The prompt used for the Subject Reference Model should focus on a single subject to achieve the most accurate results.If the prompt involves multiple subjects, additional descriptions of the specific subjects can be included. Hailuo AI can identify the corresponding facial features based on age and gender and generate the appropriate subject accordingly.
How to accurately execute prompts with multiple subjects？
- Adding gender information of the face in the prompt can help Hailuo AI lock onto the correct subject and generate the expected scene.
Facial Photo 7
Generated Video 7
Prompt 7
In the scene, a black-haired man wearing a suit sits at a European-style dining table, looking at the camera. Next to him is a woman with brown hair, smiling and gazing at the camera.
Facial Photo 8
Generated Video 8
Prompt 8
In the scene, a black-haired man in a suit sits at a European-style dining table, looking at the camera. Beside him is a woman with brown hair, smiling and gazing at the camera.
- Adding age information of the face in the prompt can help Hailuo AI lock onto the correct subject and generate the expected scene.
Facial Photo 9
Generated Video 9
Prompt 9
An elderly man in a shirt is holding a young girl in a red dress. They are sitting together in a cozy indoor setting, embracing and smiling at each other.
Facial Photo 10
Generated Video 10
Prompt 10
An elderly man in a shirt is holding a young girl in a red dress. They are sitting together in a warm, cozy indoor setting, embracing each other and smiling.