SceneCraftOpen studio ↗
TEXT TO VIDEO

AI video generator.
One scene at a time.

SceneCraft is an online AI video generator for turning a written scene into a short clip. Describe what moves, how the camera behaves and the mood you want, then choose the available video settings in the web studio.

Open AI video studio

Sign-in and web credits are required to generate. Check live availability in the studio.

SceneCraft inspiration: a night drive through a glowing city
SceneCraft brand inspiration. This artwork is not a result of the example prompt.

How to turn text into a short video

  1. Write one clear scene. Describe the subject and a single action. Add camera direction and lighting that support the movement.
  2. Choose the clip settings. Select an available model, resolution, duration and landscape or portrait frame. Review the total credit quote before submitting.
  3. Generate and review. Sign in and submit your scene. Follow the task status in the studio and revisit the approved result in your private library.

A prompt to start from

A lone cyclist follows a quiet coastal road at sunrise. The camera tracks slowly from the side while the sea glows in soft morning light. One continuous shot.

A single action and one camera movement give the clip a focused direction. Avoid asking a short generation to cover several locations or a complete story. Results vary; complex motion may need another prompt.

Models, settings and limits

Current video model: Wan 2.6. Models labeled Preview are not yet available for generation. Check the studio for the current selection.

Wan 2.6 offers 720P / 1080P resolution and 16:9 / 9:16 framing, with 5, 10, 15-second clips and silent output. MiniMax H3 photo-to-video is a separate workflow: 768P, five seconds, silent, automatic frame and an AI watermark at 25 credits per clip. Check live studio status before purchasing.

What does a video cost?

A Wan 2.6 720P 5-second silent video uses 100 credits with these settings. Resolution, duration and model choice affect the total. The studio shows the quote before submission.

Current video models and credit costs
ModelResolutionDuration and soundCredits per task
Wan 2.6720P5 seconds · silent100
Wan 2.6720P10 seconds · silent200
Wan 2.6720P15 seconds · silent300
Wan 2.61080P5 seconds · silent160
Wan 2.61080P10 seconds · silent320
Wan 2.61080P15 seconds · silent480

MiniMax H3 inputs and credit costs

Choose text-to-video, first and/or last frame, or all-in-one reference. Combine up to 9 images, 3 videos and 3 audio files, with 12 files total. Audio/video references are 2–15 seconds each, with a combined 15-second limit per type. Audio needs an image or video. Use <Picture 1>, <Video 1> and <Audio 1> in your prompt to direct editing, motion, effects, speech or language.

Output duration: 4–15 seconds. Optional native sound and prompt enhancement are included. An AI watermark is applied. 480P and 768P are native output; higher settings use post-generation upscaling.

H3 output credits, before reference processing
ResolutionCredits / second5-second example
480P315
768P525
1080P · upscaled1050
2K · upscaled1050
4K · upscaled1260

SceneCraft processing: the first image is included, each additional image adds 2 credits; video adds 3 credits per started second and audio adds 1 credit per started second, each type’s total rounded up. These are SceneCraft service charges; supplier reference input is currently free. The studio displays the combined quote. If upscaling returns 768P, the output-resolution difference is returned automatically.

Credit packs are one-time purchases, starting at $9.99 USD for 500 credits. Compare packs and generation examples.

Need a more specific starting point? Explore three video prompt examples and open one in the studio.

Before you create

Can I animate an existing image?

Use MiniMax H3 in the photo-to-video studio for a 768P, five-second silent clip at 25 credits. It chooses the frame automatically and includes an AI watermark. Check live studio status before purchasing.

Are my creations public?

Your saved library is private by default. Prompts and media are processed to provide generation and private delivery. See the privacy notice and storage and deletion policy for access and retention details.

What happens if generation fails?

Accepted tasks reserve credits. When a task is confirmed to have failed, the reserved credits are returned once. An uncertain provider outcome remains under review until reconciled. This is separate from a refund of a credit-pack purchase.

Can I use the output commercially?

Usage depends on the selected model terms, your rights to the inputs and applicable law. SceneCraft does not promise exclusive copyright or unrestricted commercial rights. Read the service terms before using an output.

Give your next scene a direction.

Create an AI video