grok / image-to-video
Make a still image move with xAI Grok — image-to-video on the cloud
- images
- api_config
- video
- workflow_id
- raw_json
The most useful of the three grok video nodes. CivitaiVideoGenGrokImageToVideo runs xAI's Grok image-to-video model on Civitai's cloud: you feed it an image (optionally several), write a prompt describing the motion, and get a clip back where the still actually moves - camera drift, action, physics. Grok's video weights aren't downloadable, so this cloud node is the only door into that model from ComfyUI, and it's billed in Buzz.
It lives in Civitai/Video/grok, part of Civitai's official ComfyUI pack. The image-to-video mode is the sweet spot of the family because it pairs perfectly with the rest of your local graph: generate a still with your local model, then hand it to grok to animate. You get the controllability of a local image and the motion of a closed video model, all in one workflow.
How it works
The node submits a videoGen workflow (engine: grok, operation: image-to-video) to Civitai's Orchestration API. Wire images into the images input (an IMAGE socket - a batch of images, e.g. from a Load Image or your local generation output), the pack uploads them, and the cloud animates them per your prompt. The result comes back as a native VIDEO tensor plus the usual workflow_id and raw_json.
The inputs that matter
- prompt (required) - what should happen in the clip. Describe motion explicitly: "camera slowly pushes in as the character turns and waves" beats "nice video."
- images - the still(s) to animate. Optional and singular-to-multiple; the
IMAGEsocket accepts a batch. - duration - seconds, 1–15, default 6.
- resolution -
480por720p, default 720p. - aspect_ratio - defaults to
autohere (it inherits from the source image), with the usual presets (16:9,9:16,1:1, …) to override.
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(justrequests).
You need a Civitai account with Buzz and credentials - a Civitai Auth node, CIVITAI_API_TOKEN for headless, or a stored key via the Civitai sidebar.
Where people get burned
- No credentials → "No Civitai credentials" error. Set
CIVITAI_API_TOKENand restart. - Your images leave the machine. This is a cloud call by design: source images and prompt go to Civitai's servers and Grok's model, subject to both parties' moderation. Don't feed it material you wouldn't upload anywhere.
- Video is the expensive meter. Per-second cloud video burns Buzz, and the community has been vocal about API video pricing. Check the node's cost report before running a batch.
- Grok refuses at the source. It's a closed model - its own content filter applies and there's no local way around it.
- Early preview. The README warns of breaking changes without notice; early community reports flagged slow jobs and bugs. Long clips can hit the default 30-minute timeout - raise it via the Auth node or
CIVITAI_COMFY_TIMEOUT.
If you've got a strong still and you want it to live, this is the node. Local image in, cloud motion out - the cleanest closed-model/video integration in the pack.
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | — | |
| durationopt | INT | 61–15 | — |
| resolutionopt | COMBO | 720p | 2 options: 480p, 720p |
| aspect_ratioopt | COMBO | auto | 8 options: auto, 16:9, 4:3, 3:2, 1:1, 2:3, +2 |
| imagesopt | IMAGE | — | |
| api_configopt | CIVITAI_CONFIG | Optional Civitai Auth connection; defaults to CIVITAI_API_TOKEN or stored OAuth login. |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| video | VIDEO | — |
| workflow_id | STRING | — |
| raw_json | STRING | — |