A folded paper boat drifts slowly across a shallow puddle, fixed low camera, soft reflections and light rain.
Use this promptMODEL & WORKFLOW GUIDE · WAN
Wan 2.1 Online Generator
Wan 2.1 supports text-to-video and first-frame image-to-video in this catalog. Choose 480P or 720P and a five- or ten-second duration before reviewing the video-credit estimate.
Open Wan 2.1Sign in and review your credit balance and the task estimate before AI generation.
Editorial stock photograph · not a model output
When to Choose Wan 2.1
A simple motion study from either a written scene or a prepared frame, with a small set of duration and resolution choices.
Wan 2.1 Inputs, Sizes and Variants
These are the settings exposed by Tintmill's connected catalog. The available duration can depend on the selected size; the studio shows compatible combinations. A listed option is not a claim about open weights, a shared architecture or identical capabilities across variants.
| Option | Input workflow | Output sizes | Clip durations |
|---|---|---|---|
| Wan 2.1 | Text to video / Image to video | 480P / 720P | 5 / 10 seconds |
A Useful Wan 2.1 Prompt
Give the shot one subject action. Use a fixed camera when the movement of the subject is the important part of the brief.
Prepare your promptWan 2.1 Example Prompts
Original editorial starting points. These prompts have not been submitted as model-quality tests.
Animate the first frame with steam rising from the cup; keep the table, background and camera still.
Use this promptA wide shot of tall grass bending once in a gentle breeze, overcast light, restrained movement, no cuts.
Use this promptVersion, Provider and Usage Rights
Origin and release
Wan-AI's official Wan2.1-T2V-14B release. Official release reference.
Architecture and workflow scope
The upstream text-to-video model combines a diffusion transformer with a causal video VAE. This is distinct from selecting an image-to-video endpoint, which also needs an opening frame.
License for the identified version
The published upstream model card identifies Apache 2.0. Hosted variants and adapters still require their own terms review.
Tintmill series names identify selected hosted models and output settings. They do not mean that Tintmill independently trained the underlying model. Review the applicable model license and service terms for your intended use. Content rights and the generation data policy explain the service relationship.
Wan 2.1 Questions
01Which Wan 2.1 options can I use here?+
This guide covers Wan 2.1. The comparison lists each option's supported inputs, sizes and durations. Availability and the credit quote are checked before submission.
02How is a Wan 2.1 request charged?+
Image tasks use image units and video tasks use a separate video-credit balance. The selected variant, output size, count and duration determine the total shown in the studio. See Pricing for purchase availability. Generation requires sufficient account credits.
03Are the Wan 2.1 examples actual generated outputs?+
The prompts are original editorial starting points, and the illustration is stock photography. Neither is a benchmark or a demonstration of this model's output. Your uploaded files and generated results remain private to your account.