视频生成 (Video)
AI-generatedSummary
Generate a video using the selected AI provider's video generation model.
Inputs
- videoPrompt (required) — Prompt describing the video to generate.
- provider — AI provider. Non-custom modes connect directly to the provider's official API. Defaults to zhipu if omitted.
- model — Model name for video generation. Defaults to 'cogvideox-flash' if not specified (per source runtime behavior).
- advancedOptions — Optional advanced parameters: temperature (0-2), maxTokens, topP, overrideApiKey, overrideApiBase, actualModel for model mapping.
Output shape
JSON object returned from the video generation API. If the response follows a standard format with 'choices', convenient fields (content, output, role) are extracted; otherwise the full raw API response is returned.
The node makes a POST request to
${baseUrl}/videos/generationswith body{ model, prompt }. API key is supplied via Bearer header for direct connections or included in the body for gateway mode.
Examples
Example 1: Generate a short video clip via Zhipu GLM
Provider: zhipu, Model: glm-4-flash (or cogvideox-flash), Video prompt: 'A sunset over the ocean', Advanced: temperature 0.7