X AI Grok Imagine Video Text To Video (Channel Edition)

The text-to-video end of Grok Imagine, channel edition. Write a prompt, get a video - 6 to 30 seconds of it, starting at 6. No image input, no ControlNet, no AnimateDiff-style motion modules, no VRAM stack. The prompt alone is the whole workflow, because the model doing the work lives on SiliconFlow's BizyAir channel, not in your ComfyUI install.

You reach for this when you want a video from a sentence and you don't want to become a video-generation specialist to get it. Local video is a genuinely different beast from local image gen - temporal consistency, frame counts, motion coherence - and the hosted models have spent their budgets making the text-to-video path feel trivial in comparison. Grok Imagine's T2V is competitive in that top photorealistic tier, and the channel edition means you don't even need an xAI account; your SiliconFlow key covers it.

How it works

Standard BizyTRD mechanics. BizyAirPlus ON, prompt in, request to the channel, VIDEO tensor out. The interesting part is the control set: duration is an integer slider from 6 to 30 seconds (default 6), which is more granular than the fixed-duration combos on the official editions. resolution defaults to 720p and aspect_ratio to 16:9 - both combos. api_config overrides base_url/api_key per request, and skip_error outputs a placeholder instead of aborting the workflow on failure. That's the whole input surface, which is honestly the point: this node is simple on purpose.

Inputs and outputs

• prompt (required, multiline) - the entire creative input.
• duration (required, 6–30s, default 6) - clip length.
• resolution (required, default 720p) and aspect_ratio (required, default 16:9).
• Outputs: video (VIDEO), request_id, urls.

Installing BizyAirPlus

Install the pack once, use it for every node in the family:

cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/siliconflow/BizyAirPlus.git
cd BizyAirPlus
python -m pip install -r requirements.txt

or search "BizyAirPlus" in ComfyUI Manager. Restart, create an account at bizyair.ai, get an API key, flip the BizyAirPlus button ON. Only dependencies: bizyair-cloudberry and bizytrd.

Common issues

The BizyAirPlus standard three: button missing means deps in the wrong environment (pip show bizyair-cloudberry), cloud failures mean key/network with details in the console, and everything is metered. One more worth stating plainly: a 30-second T2V clip costs meaningfully more than a 6-second one, and text-to-video is the most expensive thing in this pack per generation. Iterate your prompt on the default 6 seconds, then extend once you're happy.