ByteDance Seedance 2.0 Fast Text To Video Fast (Channel Edition)

Pure text-to-video is the most approachable mode in the Seedance 2.0 lineup, and the Fast tier makes it the most approachable of all: type a prompt, pick resolution, choose a duration, and a clip comes back. No reference images to curate, no first-frame to finesse - just the text. This Channel Edition node is the one to start with if you've never touched a video-generation node, or the one to burn through ideas on before moving to a reference-driven or first-last-frame node for controlled work.

How it works

Cloud execution through BizyAir: your prompt posts to api.bizyair.ai, the request hits the seedance-2-0-fast-base/text-to-video endpoint, and the finished clip downloads back into ComfyUI as a native VIDEO tensor. No local video model, no VRAM, no checkpoint downloads. Metered billing on BizyAir credits, and your prompt is processed remotely.

Inputs that matter

• prompt - the whole creative brief. Motion and camera language work best: "slow dolly-in on a neon market stall at night, rain on the awnings, steam rising." That's worth more than a paragraph of static scene description.
• resolution - defaults to 2k; native 480p/720p with 1080p/2K/4K as 720p-based upscales. (This tier's native ceiling is 720p, so the top resolutions are always upscales.)
• duration - seconds, default 8, range 4–15.
• ratio - aspect ratio, default 16:9.
• seed - integer, default 0, to reproduce a clip.
• generate_audio - default on for sound; off for silent drafts.
• return_last_frame - default off; on to get the final frame as an image, handy for feeding the next stage.
• web_search - default on; lets the model pull real-world context for the prompt. Turn it off for strictly model-internal output.
• api_config / skip_error - per-request key/base override and placeholder-on-error.

Outputs: video (into preview/save), request_id, urls.

Install and get going

ComfyUI Manager → "BizyAirPlus", or:

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

Restart, register at bizyair.ai, set the key in Settings > BizyAirPlus > API Key (or BIZYAIR_API_KEY), and click the BizyAirPlus action-bar toggle ON - no node in the pack runs while that switch is off.

Gotchas

The eternal setup trio - missing key, toggle off, deps in the wrong Python (python -m pip show bizyair-cloudberry). For pure T2V, the recurring failure is asking for too much in one clip: text-only prompts have no visual anchor, so elaborate multi-scene requests come back muddled. Keep one scene, one motion, one camera move per job. And since it's billed, use the Fast tier for exploration - this node exists precisely so you don't pay the full model's price to find out a concept doesn't work.