video-extend

To extend a short video clip without losing subject identity, route the source video URL and a continuation prompt to the Veo 3-1 extend-video endpoint via the RunComfy CLI. This produces a longer, continuous result with consistent framing and lighting.

The best way to chain shot-by-shot narratives is prompt-driven video continuation. Repeatedly extend the prior output clip by providing a new prompt describing what happens next, building a longer narrative sequence without losing the original subject identity.

Choose between fast and full-quality endpoints based on your iteration needs. Use Veo 3-1 Fast Extend for faster, lower-cost iterations during drafting, and switch to the default Veo 3-1 Extend endpoint for final, full-quality continuous video output.

Yes, you can continue an existing video past per-call duration caps by routing the source clip to the Veo 3-1 extend-video endpoints. This mechanism bypasses single-generation limits by using the existing clip as a base for continuation.

To start extending a Veo clip, you need to provide the source video URL and a clear prompt describing what should happen next. The CLI processes these inputs and writes the resulting longer clip into your chosen output directory.