Video Creator
Generate AI videos from text prompts or images — transitions, effects, and short films.
Overview
Video Creator lets you generate short AI videos from text descriptions or existing images, and chain them together into multi-clip productions using the Scene Builder. It covers two primary workflows: text-to-video for generating scenes purely from description, and image-to-video for bringing a still image to life with motion.
The resulting clips can be used as standalone social media content, embedded in presentations, used as b-roll for longer productions, or assembled into a short film using the built-in narrative tool.
Screen previews
A quick look at the screens. Use the toggle in the corner of each preview to switch between the dark and light themes.
Text to video prompt panel
Scene builder timeline
Short film scene list
AI effects library
AI transition between two images
Video Creator timeline and generation queue
Text to Video
How generation runs
Open the Text to Video tool in Video Creator, describe the scene (up to 2,000 characters), choose 5, 8 or 10 seconds and an aspect ratio (16:9, 9:16 or 1:1), then click Generate video. Video generation is asynchronous: the job is queued, checked every 5 seconds while a live panel shows the status and elapsed time, and the finished clip is played in place and can be downloaded as .mp4. You can cancel while it runs; a failed job shows its error message. Generation runs through the server on the OpenRouter video API (default model x-ai/grok-imagine-video-1.5-lite) and each clip is billed by OpenRouter.
Type a scene description in the prompt field and the AI generates a short video clip. Parameters to configure:
- Prompt — describe the scene, subject, action, setting, and mood. More detail produces better motion consistency. Example: "A solitary lighthouse on a rocky cliff, waves crashing below, dramatic storm clouds moving in fast motion, cinematic wide shot".
- Duration — choose 3, 5, or 10 seconds. Shorter clips generate faster; longer clips allow more complex motion.
- Motion Style — cinematic (smooth panning and tracking), animation (stylised movement), timelapse (sped-up environmental changes), slow motion (stretched detail shots), or documentary (handheld-style natural movement).
- Aspect Ratio — 16:9 (landscape), 9:16 (vertical/stories), 1:1 (square).
Camera Language in Prompts
Use camera movement vocabulary for precise control: "slow dolly forward", "overhead aerial pull-back", "tracking shot following the subject", "static wide shot". These terms are well understood by the AI model and produce markedly more cinematic output than simple scene descriptions alone.
Image to Video
How generation runs
Open the Image to Video tool in Video Creator. Drop a .jpg or .png image (up to 15 MB) onto the drop zone or click it to choose one; the picture becomes the first frame of the clip. Describe the motion in the prompt (for example "camera panning left, fire burning slowly"), pick a model and a duration of 5, 8 or 10 seconds, then click Generate video. The job is queued and checked every 5 seconds while a live panel shows progress and elapsed time. When it completes, the original image and the generated clip are shown side by side, and the clip can be downloaded as .mp4. The image is downscaled to at most 1280 px before upload. Generation runs through the server on the OpenRouter video API (default model bytedance/seedance-2.0-fast, also x-ai/grok-imagine-video-1.5-lite); each clip is billed by OpenRouter and needs account credits.
Upload a still image (or use one generated in Image Creator) and describe the motion you want applied to it. The AI animates the image while preserving its visual character.
The motion prompt should describe what moves and how, not what the scene looks like — the AI already knows the scene from the image. Examples:
- "The trees sway gently in the breeze, sunlight flickering through the leaves"
- "Camera slowly pans right revealing more of the landscape"
- "The waterfall flows down the rocks, mist rising at the base"
- "The character turns their head slowly toward the camera and smiles"
Supported input formats: PNG, JPEG, WebP. Maximum file size: 10MB. For best results, use high-resolution source images (at least 1024px on the shortest side).
Scene Builder
The Scene Builder chains multiple video clips into a continuous sequence with transitions between them. This is the tool for producing a complete piece of content — a product demo, a brand video, or a story — rather than a single standalone clip.
Building a Sequence
- Click New Scene Builder Project to open the timeline editor.
- Add the first clip — generate a new text-to-video clip or import an existing clip from your history.
- Click the + button after the clip to add a transition, then add the next clip.
- Continue adding clips and transitions until the sequence is complete.
- Preview the full sequence, make adjustments, then export.
Transitions
Between any two clips, select a transition effect. Available transitions include: Cut (no transition), Cross Dissolve, Fade to Black, Wipe, Zoom In/Out, and Motion Blur. Each transition can have its duration set independently (0.5s to 2s).
Short Film Creator
The Short Film Creator is a higher-level narrative tool built on top of the Scene Builder. Instead of building a sequence clip-by-clip, you write a short script or scene breakdown and the AI generates all the clips and assembles them automatically.
Write your script in the script panel — one scene per paragraph. Each paragraph should describe: location, time of day, characters present, action, and mood. The AI generates one or more clips per scene and connects them with appropriate transitions.
Script Format
You do not need formal screenwriting format. Plain paragraphs work well. The key is to be specific about the visual elements and movement in each scene. Abstract emotional descriptions like "a feeling of loneliness" need to be anchored to concrete visual action: "a lone figure sitting on a park bench watching other people pass by".
Video Effects and Colour Grading
After a clip is generated, apply post-processing effects in the Effects panel:
- Colour Grading Presets — Warm, Cool, Cinematic Teal & Orange, Desaturated, High Contrast, Vintage Film.
- Filters — Film Grain, Vignette, Lens Flare, Chromatic Aberration (subtle), Bloom.
- Speed — set the playback speed of the clip (0.25× to 4×) for slow-motion or timelapse effects in post.
- Stabilisation — AI-based stabilisation for handheld-style clips.
Effects are applied non-destructively — you can remove or adjust them at any time without regenerating the clip.