Marey Realism V1.5 | Video to Video

moonvalley/marey/pose-transfer

Ideal for matching human movement. Your input video determines human poses, gestures, and body movements that will appear in the generated video.

Inference

Commercial use

Schema

LLMs

Playground API

Input

Prompt*

Detailed Description: A venerable tribal chief, his weathered face marked with dark, ritualistic paint, stands proudly against a jungle backdrop. His elaborate headdress, a magnificent creation of numerous feathers, beads, and a central polished stone, sways gently with his movements. His initial stern expression softens into a confident smile as he begins to speak, his lips moving with unspoken words of wisdom or command. In a single, fluid motion, he raises his hand and gives a decisive wave, a gesture of both greeting and authority that underscores his leadership role within the tribe.

Background: A dense tropical rainforest is blurred into a soft, verdant backdrop, with muted greens and browns suggesting a lush, humid environment.

Middleground: The chief is the central focus, his head crowned by the large, intricate feathered headdress that moves subtly as he speaks. His shoulders and torso are visible, adorned with traditional necklaces.

Foreground: His hand lifts into the frame, palm open, executing a single, confident wave toward the viewer before lowering again. The closest feathers of his headdress rustle with the motion.

Video Url*

Hint: Drag and drop video files from your computer, video from web pages, paste from clipboard (Ctrl/Cmd+V), or provide a URL. Accepted file types: mp4, mov, webm, m4v, gif

Additional Settings

Customize your input with more control.

Result

Idle

Waiting for your input...

Your request will cost $2.00 per video.

Logs

About Moonvalley

Moonvalley is pioneering the future of AI-powered filmmaking. Our flagship model, Marey, is the world's first commercially-safe video generation model trained exclusively on licensed data. Built to meet the standards of world-class cinematography, Marey is designed for filmmakers who demand precision in every frame—offering unmatched control, consistency, and fidelity to bring any creative vision to life.

🕺 Pose Transfer

Pose Transfer lets you transfer a human performance from one video to another. Preserve an actor's movements—even facial expressions—from a reference video and apply those in a newly generated video. Use Pose Transfer with just a prompt, or in combination with a reference image to dictate the first frame of your new video.

🛠️ Pose Transfer Inputs

Here's a breakdown of the inputs you'll need for Pose Transfer:

Prompt: This is where you'll write your prompt in the Prompt description box. Be descriptive about camera details (e.g. 35mm, handheld shot), scene or environment, subjects and action of your video. Prompts with at least 50 words provide best results.
Video URL: Marey will preserve the human performance from this video. Movements, poses, and facial expressions from this video will be applied to a net new video.
Additional Settings: There are three additional settings with Pose Transfer, including reference image URL, negative prompt, and seed value.
Reference Image URL: Apply the performance from your reference video to an image of your choosing. The reference image will be the first frame of your newly generated video. Note: a reference image is optional.
Negative Prompt: Add or remove negative prompts to help guide the model away from undesirable results.
Seed: Select your desired seed value. Seed determines the randomness and different possibilities of each generation. By default, seed will be set to '9'. Use '-1' for a random seed value.

🚀 Quick Start Guide

Follow this quick start guide for using Pose Transfer on fal. It will walk you through the inputs you'll need to start generating and how to export within a few minutes.

Create a detailed prompt in the description box. Read through Marey's prompting best practices before generating.
- ✅ Do's:
  - Aim for a minimum of 50 words of detail
  - Use our recommended prompt structure outlined in the best practices document (e.g. [Camera Movement] + [Scale/Perspective] + [Core Visual] + [Environmental Details] + [Lighting/Technical Specs])
- ❌ Don'ts:
  - Add negative prompts into your prompt description—use the Negative Prompt field mentioned below.
Add a Video URL. Click 'Choose' and use the file explorer to find your desired video. The human performance from this video will be transferred to your new generation.
- Note: Pose Transfer works for human performances. An animal performance will not be preserved with this control.
Click ‘More’ on Additional Settings to expose more inputs for your generation.
Optional: Click ‘Choose’ on Image Reference URL and pick a reference image to use as your video’s starting frame.
Include negative prompt words. Marey is pre-loaded with negative prompt words to help direct the model away from unwanted results. Add or remove words as needed.
Select a seed value. Seed determines the randomness and different possibilities of each generation. By default, seed will be set to '9'. Use '-1' for a random seed value. We'll switch this to -1 for now.
Click 'Run' to start your generation.

Review and Export

By default, the results tab will show you a visual preview of your video once processing has finished. Make sure that you're satisfied with your video.

If necessary, make adjustments to your prompt and settings and then run your video again. Running the generation again with a seed of -1 will create a new video from a random seed value.

Click the 'Download' button for an MP4 of your video once you’re satisfied with the results.

Optional: Select ‘JSON’ from the results drop down menu. This will give you a unique “url” link to copy and paste into a browser for download (an easy option to quickly share your video with others).

Marey Realism V1.5 Video to Video