Magi
3 modes on this page
Magi turns a text prompt into a video clip. Pick a mode, write the prompt, run it in the studio.
Input
FormMode
Prompt*
Aspect Ratio
Resolution
Result
PreviewMotion preview, not Magi output
$0.32 per second·~90s
What Magi is for
Magi turns a written prompt into a short video clip. Runtime is about 90 seconds and compute is $0.32 per second. Modes on this page: Text-to-Video (Distilled), Extend Video (Distilled), and Image-to-Video (Distilled). It is a clip generator, not a full training-video editor.
Magi is billed in Knowlify credits at the compute cost ($0.32 per second). A typical run is about 90 seconds. Published rate: Your request will cost $0.32 to generate one four-second video. For $1 you can run this model approximately 3 times. Additional seconds will cost $0.08 each, calculated at 24 frames per second. Additional inference steps above 16 incur a 1/16 multiplier each, such that your total cost will be multiplied x2 at 32 steps.
Pick Magi for text to video at $0.32 per second and ~90s. Compare it with Ai Avatar, Argil Avatar on the related cards below if you need a different look, speed, or input type.
How to use Magi
- 1. Upload the still Magi should start from, then write the change or motion in Input.
- 2. Run it. The prompt and file hand off to the Knowlify studio, where Magi generates.
- 3. Stay in the studio if you want that frame or clip inside a longer narrated video.
What you need
- A prompt that names the subject, camera move, and lighting.
- A source still (PNG or JPG) for the model to start from.
- Aspect Ratio (auto, 16:9, 9:16, 1:1).
- Resolution (480p, 720p).
Example prompt
a dragon coiled on a burning castle keep, embers in the wind, night street photography, neon, rain, grain, no text, no watermark, slow cinematic camera move, natural motion
Frequently asked questions
How do I run Magi?
Run Magi on Knowlify at /generate/magi. This page has 3 modes in one dropdown. Write the prompt on this page and it opens in the studio, where Magi generates a short clip. The same studio also turns documents into full narrated videos.
Where can I use Magi online?
On Knowlify (knowlify.com/generate/magi). You do not need a separate playground. Magi runs in the studio next to document-to-video.
Can I make a longer video with Magi?
Magi itself produces a short clip. On Knowlify you can take that output, or skip it and upload a document, and generate a full narrated video with a script you can edit.
What do I need to run Magi?
A prompt and a source still. Magi will not invent the starting frame for you on this mode.
How long does Magi take?
Plan on about 90 seconds for a typical run, then whatever queue time the studio has. Video models sit at the high end of that range.
Does Magi have more than one mode?
Yes. This page has 3 modes: Text-to-Video (Distilled), Extend Video (Distilled), and Image-to-Video (Distilled). Switch them in the Mode dropdown instead of opening a new URL.
What does MAGI-1 (Distilled) cost?
Your request will cost $0.32 to generate one four-second video. For $1 you can run this model approximately 3 times. Additional seconds will cost $0.08 each, calculated at 24 frames per second. Additional inference steps above 16 incur a 1/16 multiplier each, such that your total cost will be multiplied x2 at 32 steps. Knowlify charges this as credits, at the compute cost.
What is MAGI-1 (Distilled) best for?
MAGI-1 distilled is a faster video generation model with exceptional understanding of physical interactions and cinematic prompts. It turns a text prompt into a video clip.
Do I need an account to use MAGI-1 (Distilled)?
Yes, a free account. The prompt you write here is waiting for you when MAGI-1 (Distilled) opens in the studio.