Create brand videos for campaigns directly on the canvas.

Google’s latest video generation model produces high-fidelity, 8-second cinematic videos with synchronized native audio and intense realism. It supports video extension, multi-shot control, guide-by-reference images, and fine-grained narrative edits. Core strengths: true-to-life textures, physics, and dialogue generation for professional storytelling.

Sora 2 is OpenAI's premium video generator and is one of the leading models in the industry. It offers audio with crisp narration and sound effects as well as realistic dynamics and motion.

OpenAI’s Sora 2 delivers highly realistic, dynamic videos with advanced physical simulation, synchronized sound, and multi-character consistency. It's built for cinematic visual effects, scene fidelity, and complex camera and movement direction—ideal for viral, artistic, and professional content.

If you are using the Infinite Canvas, place an image on the canvas and select it. You will see the Image Menu appear at the bottom of the screen. On this look for a green Video icon. Click on this and it will open up the StyleForge Video Generator model. You can select from the different available models that support the size of image you have selected. Each model has its own aspect ratio requirements, so we filter out the models that do not offer a similar aspect ratio that matches the image you want to make into a video.

Runway’s fastest model generates 10-second, dynamic, high-resolution cinematic clips in just 30 seconds, excelling at smooth animation, real-world physics, and prompt adherence. Supports creative prototyping, quick iterations, and real-time exploration for creators and filmmakers.

Transforms a single image and motion/audio signals into lifelike, full-body videos; handles challenging human gestures, natural speech, expression mapping, and lip-sync. Adapts freely to portrait, half-body, cartoon, and stylized outputs. Ideal for realistic avatar, performer, or influencer cloning scenarios.

Once you have selected the image on your canvas, click on the Make Video From Image button on the Image Menu (green video icon). This opens up the StyleForge Video Generator. Select the model you want. In this example, if you select Veo 3, you will see four steps, step two shows you the image to confirm that is what you want to convert to a video, step three shows you the options to enter in either a Natural Language prompt or select a more structured JSON prompt. JSON prompts are exclusive to the Veo 3 model and have shown that they enhance video output adherence by providing the model with a structured set of instructions. This is a more advanced feature and is covered in our tutorials and help systems. The fourth step is where you review the settings and click the Generate Video button. Leave the model open and wait for the video to generate (2 mintutes or so). The returning video will show in the model and you can preview and download it. It is also saved to the canvas and will appear next to the original image it was made from. You can click on the video on your canvas and the Video Menu shows you options like converting it into an AI Avatar, or previewing it, downloading it, or saving it to a brand kit. You can also save videos to albums you make for longer term easy access.

Animates characters by transferring movement, expressions, and speech from driving videos to any reference image or character. It features advanced full-body motion capture, expressive lip sync, and studio-quality facial tracking for accessible professional animation workflows.

Yes, if you are working with a video canvas project, we have multiple agentic driven workflows such as the Create AI Avatar Video, Create Green Screen Avatar Video, and the B-Roll Footage you can open up to generate up to 6 videos at the same time. Each of these agentic workflows support vision analysis and script / action sequence inputs. Some models support both image-to-video and text-to-video (Sora 2). The returning videos can be batch added to a brand kit, batch downloaded, or added to a video canvas scene you are making. Our batch processing for videos currently supports both the Veo 3.1 and Sora 2 video models.

Seedance 1 Pro unifies text-to-video, image-to-video, and editing within one model; enables professional-grade macro videos, creative compositing, and advanced multimodal workflows. It excels at fast high-definition output, style transfer, frame consistency, and cross-modal reasoning. Especially fit for campaign, social, and product work.

Maps natural speech and emotional expressions onto AI-generated or uploaded video with precision; supports voice-over alignment, background music, and professional dubbing workflows. Empowers realistic talking avatars and character-driven storytelling.

Generates realistic, motion-rich videos from text or images with synchronized audio, robust expression, and accurate scene context. Supports instruction-based editing, visual reasoning, and multilingual dialogue for streamlined, cinematic content creation.

Kling Master is known for superb dynamics and prompt adherence. It can generate 5s and 10s long videos at 1080p quality.

Kling 2.1 Master is known for superb dynamics and prompt adherence and can create 5s and 10s generations at 1080p quality.

Wan 2.5 Fast by Alibaba is an affordable text-to-video generator that supports audio and generates both 5s and 10s long videos at 1280*720p.

A premium text-to-video engine, delivering ultra-realistic motion, cinematic lighting, and fluid camera transitions for advanced narrative control. Supports multi-character complexity, high prompt fidelity, 1080–4K output, and scene conditioning for studio-level video demand.