DeepMode Blogs | Blogs about AI character generation

Video LoRAs are small modules trained on specific motion or visual styles.

They work like image LoRAs but for video, allowing you to apply effects like "underwater" or "camera kiss" without retraining the whole model.

Some LoRAs are optimized for Text-to-Video (T2V) prompts, while others shine best when used with Image-to-Video (I2V) setups.

What's the Difference?

Text-to-Video (T2V): You start with a text prompt. DeepMode generates both the initial frame and the animation based on your description. This is ideal for full-scene generation where everything is guided by your prompt and selected LoRA.

Image-to-Video (I2V): You start with an image. DeepMode animates that image, keeping the character or composition mostly intact while applying movement or style from your selected Video LoRA. This is great for refining single-image art into moving clips. 

Remember for Image-to-Video (I2V), it is recommended to use a good starting image that will match the action of the LORA.

This article will help guide you generate videos using LORAs and understand which one works best for I2V and T2V! If you are looking for NSFW Video LORA guide, you can click here to read the blog!

Lets Dive In!

1. Hero Run Effect

Trigger Words: 4ct3ion Action Run

Recommended use: Image-to-Video (I2V) / Text-to-Video (T2V)

This LORAs is about mimicking that iconic high-speed "hero run" effect seen in action films and superhero scenes. It gives your model that cinematic scene where he/she runs towards the camera in a dramatic shot.

While it also works in text to video, you can get better and more stable results from image to video. 

Prompt example:

T2V

LORA Weight: 1.0

Man in street running towards the camera,. 4ct3ion Action Run, camera is following him. In his background are exploding cars and guns blazing, 4ct3ion Action Run

I2V

LORA Weight: 1.0

woman, run dramatically towards the camera 4ct3ion Action Run

2. Zoom In Camera Motion

Trigger Words: cr34sh crash zoom in effect

Recommended use: Image-to-Video (I2V) / Text-to-Video (T2V)

This adds a smooth, dramatic zoom-in effect to your video, perfect for focusing attention on a character or building intensity in a scene.  While it is intended for I2V, you can also generate cool stuff using T2V. 

Prompt example:

T2V

LORA Weight: 1.2

A woman, wearing a t-shirt and jeans, sits in a wooden chair, facing away from the camera. She is outdoor. The camera performs a cr34sh crash zoom in effect, rapidly zooming closer to the woman's face as she turns her head, looking directly at the viewer then she smiles and waves.

I2V

LORA Weight: 1.0

Woman in bed. The camera performs a cr34sh crash zoom in effect. The camera zooms rapidly to her face and then she smiles.

3. Robotic FaceReveal

Trigger words: r8b8t1c robotic face reveal

Recommended use: Image-to-Video (I2V)

A lora that transforms a human face into a robotic one mid-scene with a clean, cinematic peel-back or split effect. Great for sci-fi or cyberpunk themes.

This is recommended for image to video only.

Prompt example:

I2V

LORA Weight: 1.0

A young man with dark curly hair and a neutral expression. The r8b8t1c robotic face reveal commences, lines forming as a metallic mask replaces his features.

I2V

LORA Weight: 1.0

woman walking towards the camera, zooming to her face while revealing her robotic face. r8b8t1c robotic face reveal

4. Turn to ashes

Trigger words: flian3 special effects

Recommended use: Text-to-Video (T2V)

Creates a powerful disintegration effect where the character dissolves into dust or ash - dramatic, emotional, and highly cinematic. Like the Thanos snap. 

This lora is for Text-to-Video (T2V).

Prompt example:

I2V

LORA Weight: 1.3

A man shoots a gun, looks at the camera with concern face and then he turns into ashes., flian3 special effects

TIP: You may need to increase the lora weight for this one up to 1.5 to create the ash effect.

5. Camera Kiss

Trigger Words: In the first-person perspective, the girl's lips press close to the lens as if to kiss you - her lips fill the entire screen, Camera zooms and she kisses the camera

Recommended use: Image-to-Video (I2V)

Generates a close-up kissing animation toward the "camera," giving a POV-style moment with intimacy and expression. Using this on T2V doesn't generate good results. But when it comes to I2V, you'll love the outcome! 

Prompt example:

I2V

LORA Weight: 1.0

woman, In the first-person perspective, the girl's lips press close to the lens as if to kiss you - her lips fill the entire screen, Camera zooms and she kisses the camera

I2V

LORA Weight: 0.8

A woman looking at the camera. It zooms in. In the first-person perspective, the girl's lips press close to the lens as if to kiss you - her lips fill the entire screen.

6. Underwater

Trigger Words: woman swimming underwater, tiny bubbles coming out of her mouth and nose

Recommended use: Text-to-Video (I2V)

This is a LoRA that enhances the look of underwater scenes. Applies underwater visuals like hair floating, soft lighting, and ambient bubbles .  Prompt example:

T2V

LORA Weight: 1.0

woman underwater in a swimming pool, she is wearing a swimsuit, she is holding her breath sitting cross legged on the pool's bottom. tiny bubbles coming out of her mouth and nose. Her neoprene shorts are yellow with a black stripe on the side. Sunlight filters through the water surface above her, casting soft light rays and caustics on her body and the pool floor around her.

T2V

LORA Weight: 1.0

A woman swimming underwater, she is wearing bikini. She is inside an underwater cave, her body parallel to the cave's wall as she effortlessly glides through the water with long strokes of her toned legs, the camera follows her from behind as she swims away. A few air bubbles are trailing behind her. Black diving fins. She swims out of the cave through a narrow opening on the roof. Her neoprene shorts are yellow with a black stripe on the side. Sunlight filters through the slightly murky, dark water, casting soft light rays.*

7. Arc Shot Camera Motion

Trigger 34Ar2c arc the camera moves in a smooth curve around

Recommended use: Text-to-Video (I2V)

Create an arc motion camera that focuses on your model as it rotates around it. 

Prompt example:

I2V

LORA Weight: 1.0

Woman outdoor, it is sunny and she is looking at the camera. 34Ar2c arc the camera moves in a smooth curve around.

Don't have video models? No worries. We have a public video model that you can use to generate videos!

If you are Interested to know about NSFW Video LORAs, check out this blog.

We add LORAs from time to time. Join our Discord to see our community's awesome work and suggest some new LORAs you would like to be added!