grok / text-to-video

Type a prompt, get a video. CivitaiVideoGenGrokTextToVideo runs xAI's Grok video model on Civitai's cloud - the text-to-video flavor, which means all you need is a description. It returns a native VIDEO output you can preview, save, or feed into the rest of your graph. Grok's video weights aren't something you can download, so this is the only way it enters a ComfyUI workflow: through the cloud, billed in Buzz.

It lives in Civitai/Video/grok, part of Civitai's official ComfyUI pack. The three grok nodes are split by input - text-to-video, image-to-video, and edit-video - and this is the purest of them. Reach for it when you want a shot that exists only in your head: a establishing drone pass, a stylized motion scene, an abstract loop. No starting image, no source clip, just the prompt.

How it works

The node submits a videoGen workflow (engine: grok, operation: text-to-video) to Civitai's Orchestration API. Your prompt goes up, Civitai's fleet runs Grok's model, and the pack long-polls until the job finishes, downloads the clip, and converts it to a Comfy VIDEO tensor. The node shows its workflow_id and Buzz cost on the canvas after a run.

The inputs that matter

• prompt (required) - your description of the shot. This is the whole input, so write it like a director's note: subject, action, camera, mood.
• duration - seconds, 1–15, default 6. Longer is pricier and slower.
• resolution - 480p or 720p (default 720p). Note there's no 1080p option on the grok nodes - keep expectations aligned.
• aspect_ratio - 16:9, 4:3, 3:2, 1:1, 2:3, 3:4, or 9:16; default 16:9. Pick vertical (9:16) for shorts-style content.

Outputs: video (the clip), plus the standard workflow_id and raw_json for debugging and cost inspection.

Installing it

This is one of ~160 nodes in Civitai Comfy Nodes, Civitai's official pack for their Orchestration API:

• ComfyUI Manager: Manager → Custom Nodes Manager → search Civitai Comfy Nodes → Install, then restart.
• CLI: comfy node registry-install civitai-comfy-nodes
• Source: cd ComfyUI/custom_nodes && git clone https://github.com/civitai/civitai-comfy-nodes.git && pip install -r civitai-comfy-nodes/requirements.txt (just requests).

You need a Civitai account with Buzz and credentials - a Civitai Auth node, CIVITAI_API_TOKEN (best for headless), or a stored key via the Civitai sidebar.

Where people get burned

• No credentials → hard error. Export CIVITAI_API_TOKEN and restart ComfyUI; headless installs can't do browser OAuth.
• Video is the expensive meter. Per-second cloud video burns Buzz fast - the community's standard objection to API video nodes is exactly this cost, and it applies here too. Check the cost report on the node before running long batches.
• Grok's own filtering applies. It's a closed model with its own content moderation on top of Civitai's; the prompt can't be "abliterated" the way a local model's weights can. If the model refuses, there's no local patch.
• Early preview. The pack README warns of breaking changes without notice, and early community reports flagged slow jobs and bugs on these nodes. Generations can outlast the default 30-minute timeout - raise it via the Auth node or CIVITAI_COMFY_TIMEOUT.

For a quick "here's a scene, make it move" shot, this is the zero-friction path to Grok video inside ComfyUI. Keep an eye on the meter and treat it as a capability you can't run locally - because that's exactly what it is.