Create a short video from a prompt or animate a supplied opening frame with an optional guided ending.
Use text-to-video for prompt-led scenes, or image-to-video when the opening composition should stay anchored.
Final usage is the selected resolution rate multiplied by output duration.
Similar Models
Create videos up to 30 seconds, combine as many as 50 multimodal references, and refine specific moments with frame-level control.
A multimodal video generation model for text prompts, first-frame animation, and reference-guided production clips.
Create 4 to 10-second videos from text, up to seven reference images, or one short reference video with synchronized sound.
Model overview
FLUX.3 on sjolt supports text-to-video and image-to-video creation with whole-second durations, selectable resolution, and automatic or fixed aspect ratios through the Playground and task API.
Describe the subject, action, environment, camera motion, lighting, and visual style to turn a written direction into a short video clip.
Image-to-video requires an opening image and can also use an optional ending image to guide a transition between two compositions.
Choose any whole-second duration from 5 to 20, select 720p or 1080p, and use automatic or fixed framing for the intended placement.
Model characteristics
Turn a written brief into a complete scene by specifying the subject, action, environment, camera movement, lighting, and visual style. Duration, resolution, and aspect ratio remain independently selectable.
Use one required opening image to anchor the initial subject, composition, and visual direction. The prompt then describes how the scene, camera, and motion should develop.
Add an optional final image when the clip must arrive at a specific closing composition. sjolt automatically routes requests with both images through FLUX.3 first-and-last-frame generation.
Submit either public route through the same asynchronous task flow, receive a task identifier, then poll its status or use an optional webhook for completion updates.
FAQ
sjolt exposes text-to-video and image-to-video routes. Text-to-video accepts a prompt without frame inputs. Image-to-video requires image_url and automatically selects single-frame or first-and-last-frame generation according to whether end_image_url is present.
Choose any whole-second duration from 5 through 20, 720p or 1080p, and auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16 framing. Requests default to 5 seconds, 720p, and auto.
Yes. Image-to-video requires image_url for the opening frame and accepts optional end_image_url for the closing frame. The same prompt directs the motion and transition between those two compositions.
The Playground accepts one JPG, JPEG, PNG, or WebP opening image and one optional closing image. API requests use public HTTP(S) image URLs in image_url and end_image_url.
Describe the subject and starting state, the visible action, the environment, camera movement, lighting, and visual style. For image-to-video, focus the prompt on how the supplied opening frame should move or change over time.
A successful creation request returns a task_id. Poll the shared task-status endpoint until the task completes, or provide a top-level webhook URL to receive completion updates before retrieving the generated video.
