TanStackOSS
    ai
    feat(ai-grok): video generation adapter for the grok-imagine video models Adds a grokVideo adapter to @tanstack/ai-grok for xAI's Imagine video models (grok-imagine-video at $0.05/s, grok-imagine-video-1.5-preview at $0.08/s) using the experimental generateVideo() jobs/polling architecture: POST /v1/videos/generations to create, GET /v1/videos/{request_id} to poll, hosted mp4 URL plus usage (billed seconds + exact USD cost) on completion. Sizing follows the grok-imagine aspect-ratio template ('16:9_720p' → aspect_ratio/resolution); durations are 1-15 integer seconds; image-to-video starting frames go through modelOptions.image. The Imagine video endpoints are plain JSON (not in the OpenAI SDK), so the adapter issues direct requests with an injectable fetch seam. Closes #705. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
    nx run-many --targets=build --exclude=examples/**,testing/**
Succeeded
CI Pipeline Execution

nx run-many --targets=build --exclude=examples/**,testing/**

Click to copy
Linux
4 CPU cores
read-write access token used
dd449075742
100%

Cache hits

32 of 32 tasks used cache.

Get faster results

Learn how to enable distribution.

Atomizer enabled

32 groups of tasks optimized.

© 2026 - Nx Cloud

Terms of ServicePrivacy PolicyChangelogStatusDocsContact Nx CloudPricingCompany@NxDevTools