Zhenzhen_vidu_text2video

Vidu text-to-video is the purest version of the deal the whole Comfyui_Comfly pack makes: you describe a scene, ByteDance's model builds it, and the relay in the middle handles the account, the billing, and the polling. Comfly_vidu_text2video is that pipeline as one node - prompt in, video out - with Vidu's aspect ratios, resolutions, motion settings, and optional BGM exposed as inputs.

Like the pack's other Vidu nodes it's a paid call to ai.comfly.chat: POST to /vidu/v2/text2video, then poll the task endpoint until the clip lands. No local model, no GPU usage, and the node is patient - it'll keep checking for a good long while.

Inputs that matter

• prompt - the whole ballgame here. Text-to-video quality is prompt quality; describe subject, action, camera, and mood.
• model - viduq2 (default), viduq1, or vidu1.5. This node's list is shorter than the img2video node's, so don't go looking for viduq2-pro here.
• style - general or anime. One click for the anime look, which is Vidu's bread and butter.
• duration - 1–10 seconds.
• aspect_ratio - 16:9, 9:16, 3:4, 4:3, or 1:1. Notably more choice than the Sora node in this pack.
• resolution - 360p through 1080p.
• movement_amplitude - auto/small/medium/large.
• bgm, off_peak, watermark (+wm_position), seed - same extras as the img2video node, same semantics.

Outputs

The pack standard: video (a VIDEO object for ComfyUI's Save Video node), video_url (hosted link), task_id (track it on the relay), and response (JSON log, errors included).

Install

cd ComfyUI/custom_nodes
git clone https://github.com/ainewsto/Comfyui_Comfly

or ComfyUI Manager → search "Comfyui_Comfly" → install → restart. API key from ai.comfly.chat into the api_key field.

The honest take

This is the entry point of the Vidu family - the cheapest thing to reach for when you want to test whether Vidu's style suits your project before paying for image- or reference-driven runs. Start on viduq2, 5 seconds, 720p, general, and a well-written prompt; that's the minimal viable experiment. The anime toggle is genuinely worth trying even if you don't normally work anime - it's where Vidu overperforms. Just keep the same expectations you'd bring to any text-to-video API: a great prompt gets you a decent clip, and a mediocre one gets you a museum of smeared hands. The node won't fix the prompt, but it will faithfully bill you for every attempt.