Text to Video → Generate Video

This node has been unified into Generate Video.

The legacy text-to-video node type has been replaced by the unified Generate Video node, which drives by connection — text-only mode, image-to-video mode, first+last frame mode, or reference mode, all from one node. The provider catalog, parameter set, and credit pricing for text-to-video usage are unchanged.

Migration

Existing workflows continue to work — text-to-video nodes auto-migrate to generate-video the moment the workflow opens in the editor. Handle ids are renamed (the legacy in prompt handle becomes the typed prompt handle, cinematography becomes look / elements), and the node type is rewritten in-memory. The migration is idempotent.

What to do now

Pricing & parameters

All text-to-video pricing examples (VEO 3.x, Gemini Omni, Kling, Seedance 2, Hailuo Standard, Bytedance, MiniMax, Wan 2.6 / 2.7 / 3.0, HappyHorse, Runway, …), per-provider parameter tables, and the dispatch rules that route a wireless node to text-to-video mode are documented on the Generate Video page.

/v1/text-to-video remains a live API route and behaves identically to the unified node. A resolution, aspect ratio or duration the wired model does not support is corrected rather than rejected, and you are billed for the corrected value because it is also the value sent to the provider: off-list values snap to the nearest supported option (never the cheapest, and never landscape from a portrait request), an omitted resolution is sent at the band it is priced at, 4K canonicalises to 4k, and LTX 2.3 durations move to the nearest seeded rung. The route returns the corrections in an adjustments array. See Resolution, aspect ratio and duration corrections and the API reference on Parameter corrections.

See also

Pre & post text

This node supports promptPrefix / promptSuffix — optional text wrapped around the prompt at run time (settings panel → Pre & post text; hidden from app users; captured by presets). See Prompt pre & post text.