Wan-2.1 Image-to-Video with LoRAs

Wan-2.1 Image-to-Video with LoRAs is the Wan 2.1 adapted workflow for image-anchored animation. It models visual appearance and time together so the route’s written direction, source imagery, speech, masks, or structural conditions influence one coherent result rather than a collection of disconnected frames or a generic endpoint response.

Choose this exact route when the available assets and deliverable specifically call for image-anchored animation. Identify every subject and source, state the desired action or transformation, then describe camera behavior, pacing, environment, lighting, atmosphere, and preservation requirements; local schema fields handle resolution, duration, seed, and delivery separately from the creative brief.

Highlights

• 2.1 LoRA Reference-frame animation. Uses the supplied still as the visual foundation for motion.
• 2.1 LoRA Subject identity retention. Keeps important appearance cues recognizable as the scene develops.
• 2.1 LoRA Natural temporal movement. Creates coherent subject and environmental dynamics across frames.
• 2.1 LoRA Prompt-directed cinematography. Interprets action, camera movement, pacing, lighting, and atmosphere. The route adds a lightweight task adaptation.

Pricing

When to Use

Prompt Guide

Identify the primary subject and every source or condition, state the intended transformation or action, then describe composition, camera or viewpoint, lighting, materials, pacing, atmosphere, and exact preservation requirements. Refer to multiple inputs in their schema order.

{
  "prompt": "A precisely directed composition with explicit subject, transformation, camera or viewpoint, lighting, material, and preservation requirements",
  "image": "https://example.com/reference.png",
  "seed": 1,
  "loras": [],
  "resolution": "480p"
}

Technical Specs

Related Models

• alibaba/wan/2.1/text-to-video
• alibaba/wan/2.1/image-to-video
• alibaba/wan/2.1/vace