Nodes/comfyui-evolink/EvoLink Grok Imagine Image-to-Video (Official)
ComfyUI Node

EvoLink Grok Imagine Image-to-Video (Official)

Animate your image with xAI — and a mode knob that changes the vibe

By deeplearning-goethe·Created 24 days ago·Updated a day ago· 0
EvoLink Grok Imagine Image-to-Video (Official)
  • image1
  • image2
  • image3
  • video
  • task_id
  • status_info
  • result_urls
  • response_json
prompt
duration6
modenormal
aspect_ratio16:9
quality480p
image_urls_text
api_key
timeout_seconds1200

EvoLink_GrokVideoI2V is xAI's image-to-video node, and it has one thing nobody else in this pack's lineup does: a mode dial with a setting called spicy. Grok's image and video stack is known for running a lighter content filter than the other majors, and this node exposes that directly as three generation styles - normal, fun, and spicy - for when you want the tone of the result nudged somewhere bolder. It's the xAI personality you're paying for.

The task itself is straightforward: first image required, up to seven reference images total, and it generates a 6–30 second clip.

The inputs that matter

  • image1 (required) - the first frame / primary reference. Connect Load Image here.
  • prompt (required) - shot and motion description. The model has your subject from the images; this should be about what happens.
  • mode - normal (default), fun, or spicy. This is the fun one; expect the fun/spicy variants to be looser in both content and composition.
  • duration - 6–30 seconds, default 6, billed by duration.
  • quality - 480p (default) or 720p. Budget note: Grok's I2V tops out at 720p in this node, so don't expect 4K.
  • aspect_ratio - 16:9 (default), 9:16, 1:1, 3:2, 2:3.

image2 and image3 are optional additional references (up to seven with image_urls_text), and the pack-standard api_key + timeout_seconds (1200) round it out.

How it works

Standard pack mechanics: images upload through EvoLink's official file channel (auto-deleted after 72h, no third-party host), the job goes to POST /v1/videos/generations, and the pack polls until completion, downloading the clip as a VIDEO object. Wire video into a native Save Video node. You also get task_id, status_info, result_urls (24h expiry), response_json - the same five outputs as every video node here. Needs a recent ComfyUI for VIDEO output support.

Installing

Nothing heavy - no GPU, no model files, only requests:

cd ComfyUI/custom_nodes
git clone https://github.com/deeplearning-goethe/comfyui-evolink.git

Restart ComfyUI fully afterwards, or install through ComfyUI Manager by searching EvoLink.

Troubleshooting

Same error codes as the whole pack - 401 key, 402 balance, 403 access, 429 rate limit. VPN users: whitelist api.evolink.ai and files-api.evolink.ai. One Grok-specific thing worth knowing: spicy is relative, not a promise - xAI's moderation still runs on their servers, so if a prompt gets refused, tone it down and retry (blocked jobs don't charge). And if the clip comes back looking nothing like your reference image, check that image1 is actually connected; an empty first-frame socket changes the whole request.

CategoryEvoLink

Inputs (11)

NameTypeDefaultDescription
promptSTRING镜头/画面描述。
image1IMAGE首帧/参考图(必接,最多 7 张)
durationoptINT66–30视频时长(秒),6~30 任意整数,按时长计费
modeoptCOMBOnormal生成风格:normal 常规 / fun 趣味 / spicy 大胆
aspect_ratiooptCOMBO16:9画幅比例
qualityoptCOMBO480p分辨率
image2optIMAGE参考图 2(可选)
image3optIMAGE参考图 3(可选)
image_urls_textoptSTRING进阶:已有公网图片 URL 时每行贴一个,与 IMAGE 口混用
api_keyoptSTRINGEvoLink API Key(sk- 开头)。首次填写后自动保存到本机配置,之后可留空。分享工作流前请清空此框。获取:evolink.ai/dashboard/keys
timeout_secondsoptINT120060–3600最长等待时间;生成失败或审核拦截的任务不扣费

Outputs (5)

NameTypeDescription
videoVIDEO生成结果(已下载落地,可直连保存节点)
task_idSTRINGEvoLink 任务 ID,可到 evolink.ai/zh/dashboard/logs 查询
status_infoSTRING人读状态摘要(模型/用时/消耗 credits)
result_urlsSTRING结果原始链接,每行一个(24 小时过期)
response_jsonSTRING平台 GET /v1/tasks/{id} 的完整 JSON 响应(status/usage/results 等),供下游节点解析