Still image to video
Start from a single image and generate a short video that keeps the original frame's subject, lighting, and visual detail as the anchor.
xAI's latest image-to-video preview model transforms a single still image into a cinematic video with a written motion prompt. Direct the camera, pacing, atmosphere, and sound design from one source frame.
Prompt Direction
(0-5s) Medium shot, she speaks warmly while gesturing to a product. (5-10s) Slow push-in to a close-up, glowing skin and expressive eyes. (10-15s) Cut to an over-shoulder framing of the vanity as she smiles. Glossy, warm, cinematic.
According to xAI's preview announcement, the model is designed to convert a still image into a smooth cinematic video while following prompt-level direction for movement, camera behavior, pacing, atmosphere, and sound design.
Start from a single image and generate a short video that keeps the original frame's subject, lighting, and visual detail as the anchor.
Use the prompt to describe action, camera moves, pacing, atmosphere, and even sound-design intent so the output feels directed rather than randomly animated.
Animate storyboards, product frames, or character beats one shot at a time, then stitch the resulting clips into longer sequences in your own pipeline.
Grok Imagine 1.5 Preview is most useful when the source frame matters: a product still, a portrait, a concept frame, or a storyboard panel that needs motion without losing its starting identity.
The model requires exactly one image URL for this preview workflow.
Guide subject action, camera language, pacing, atmosphere, lighting continuity, and sound-design intent in the prompt.
APIPod queues the request and returns a task identifier while keeping xAI provider-specific naming internal.
Poll for the completed 720p clip, then repeat the process shot by shot for longer narrative or product sequences.
Use the public APIpod model ID directly. The adapter keeps provider-specific naming internal while your application depends on a stable request contract for single-image image-to-video tasks.
# Grok Imagine 1.5 Preview request payload = { "model": "grok-imagine-1.5-preview", "prompt": "Slow dolly-in on the product, soft studio light, deliberate pacing, subtle ambient sound design", "image_url": "https://cdn.example.com/ref.jpg", "duration": 10, "aspect_ratio": "16:9" } response = requests.post( "https://api.apipod.ai/v1/videos/generations", json=payload, headers={"Authorization": "Bearer $APIPOD_API_KEY"} )
The model story above follows xAI's preview announcement. The values below mirror the current validation and adapter behavior for grok-imagine-1.5-preview in APIpod.
Public model
grok-imagine-1.5-preview
Use this model ID in public API requests.
Reference images
Exactly 1
image_url is required and multiple images are rejected.
Video length
1-15 seconds
The default request length is 10 seconds when omitted.
Resolution
720p
The current TTapi-backed preview path validates 720p output.
Aspect ratios
16:9, 9:16, 1:1
3:2 and 2:3 are also accepted for this model.
Execution
Async task
Submit generation, store the task ID, and fetch status until completion.
Yes. The current APIpod validation requires image_url for grok-imagine-1.5-preview, and the adapter rejects requests without a reference image.
xAI positions Grok Imagine 1.5 Preview as a model that follows motion direction from text. In practice, write prompts that cover subject movement, camera moves, pacing, atmosphere, lighting continuity, and sound-design intent.
Yes. xAI describes a workflow where individual still frames can be animated and then combined into a longer sequence. In APIpod, you would generate each shot as its own async task and assemble the clips downstream.
The current contract accepts 1 to 15 seconds. If you omit duration, APIpod defaults this model to 10 seconds.
Use 720p for this preview model. The current adapter path validates 720p as the supported resolution.
APIPod accepts 16:9, 9:16, 1:1, 3:2, and 2:3 for grok-imagine-1.5-preview.
No. Use the public APIpod model ID: grok-imagine-1.5-preview. Provider-specific model naming is handled internally.
Commercial usage depends on your account terms, content rights, and the applicable provider policy. Check APIpod Terms of Service before production release.
Bring xAI's latest image-to-video preview model into your product workflow with one public model ID and the same APIpod generation pipeline.