AI Generator

A community-driven, open-source model delivering state-of-the-art video quality—sharp details, smooth motion, and lightning-fast generation. Open-source SOTA on benchmarks, runs on 8 GB GPUs, multi-task (T2V, I2V, editing)

1. Wan 2.1 (Core)

Sizes: 1.3 B (“Light”) & 14 B (“Pro”)

• What it does: Text→Video and Image→Video generation
• Why you’ll love it:
  • Light (1.3 B): Blazing-fast inference on an 8 GB GPU—perfect for quick demos and prototypes.
  • Pro (14 B): Industry-grade visual quality with the richest detail and smoothest motion.
• Integrations: Full Promptus ComfyUI support via Promptus app or Promptus web

2. Wan 2.1 VACE

Sizes: 1.3 B (“Light”) & 14 B (“Pro”)

• What it does:
  • Text→Video & Image→Video generation
  • Video→Video editing (full-clip)
  • Masked Video→Video (localized in-frame edits)
• Why you’ll love it:
  • Unified “Video Condition Unit” lets you steer with depth maps, poses, optical flow, masks, layouts, and more.
  • One model, endless possibilities—no need to switch checkpoints for different workflows.
• Integrations:
  • Full Promptus ComfyUI support via Promptus app or Promptus web

3. Wan 2.1 Fun Control

Size: 14 B (under the hood)

• What it does: Pose-driven animation from any static image
• Why you’ll love it:
  • Leverage a human-pose ControlNet to bring characters to life with realistic movement.
  • Ideal for animated storytelling, character tests, and rapid prototyping.
• Integrations: Full Promptus ComfyUI support via Promptus app or Promptus web

4. Wan 2.1 Fun Camera

Size: 14 B (under the hood)

• What it does: Virtual camera-path synthesis (dolly, pan, tilt, and more)
• Why you’ll love it:
  • Simulate professional cinematography moves with simple text or UI controls.
  • Create dynamic shots without a physical camera.
• Integrations: Full Promptus ComfyUI support via Promptus app or Promptus web

5. Wan 2.1 Fun InP

Size: 14 B (under the hood)

• What it does: Video in-painting & overlay (mask-based edits)
• Why you’ll love it:
  • Seamlessly replace or extend footage in specific regions.
  • Perfect for VFX touch-ups, logo placements, and seamless scene extensions.
• Integrations: Full Promptus ComfyUI support via Promptus app or Promptus web

6. Wan 2.1 First-Last Frame

Size: 14 B (under the hood)

• What it does: Long-clip generation from only first and last frames
• Why you’ll love it:
  • Auto-interpolates intermediate frames with smooth temporal consistency.
  • Enables multi-second narratives without scripting every keyframe.
• Integrations: Full Promptus ComfyUI support via Promptus app or Promptus web

Choosing Your Model

• Quick & Easy: Go Light (1.3 B) core or VACE on an 8 GB GPU for sub-second previews.
• Maximum Quality & Control: Upgrade to 14 B VACE for full editing pipelines and multi-modal inputs.
• Specialized Workflows: Use Fun Control, Fun Camera, Fun InP or First-Last Frame in Promptus ComfyUI (app or web) to unlock pose, camera, in-paint, or temporal-interpolation magic.

‍