Tokease Seedance 2.0 Video Generation

Problem Space

Video generation APIs create more work than a single request: a caller must create an async job, poll status, parse progress, download a time-sensitive video_url, and surface failures clearly. If every project assembles its own script, handling of the API Key, task_id, polling interval, and download path can drift. A fixed CLI is easier to reuse when prompts, negative prompts, aspect ratios, and durations need to repeat across jobs. This skill wraps the Seedance video workflow on tokease.cn as a developer-oriented command-line flow.

Core Capabilities and Steps

• Default model: generates AI video with the Seedance 2.0/uP model.
• Authentication: asks for a tokease.cn API Key before the call and passes it with --api-key; the source states that environment variables are no longer relied on.
• Job workflow: uses seedance.py to create a job, query its status, and download the result, while keeping the job ID available for later checks.
• Generation controls: supports --prompt, --negative-prompt, --ratio, --duration, --resolution, --seed, --wait, --download, and --interval.
• Status handling: recognizes pending, running, completed, and failed, then retrieves the video from output.video_url after completion.

Fit and Caveats

This is useful when a script, automation, or AI assistant needs to produce short videos consistently, especially when the task ID must remain queryable. The video URL is time-sensitive, so download it promptly after completion. When using --wait, the default polling interval is 15 seconds. On failure, surface the error message and first check the prompt, API Key, model, or parameter settings.