InfiniTalk

InfiniTalk is a talking head where you set the length, from a few seconds to about twenty.

InfiniTalk animates a still portrait against an audio track and differs from its neighbours in one useful way: you choose how long the result runs. That makes it easy to fill a fixed-duration slot (a five-second bumper, a ten-second card) without trimming afterwards.

The trade is that the audio has to be at least as long as the length you ask for. A shorter track is rejected before anything is charged, so pick the length that fits the slot and give it a track that covers it.

At a glance

  • Type: Image + audio → video
  • Inputs: Still portrait, audio track, prompt
  • Length: Chosen: ~5, 10, 14 or 19 seconds
  • Resolution: 480p · 720p
  • Pricing: Per second of output

What it is best at

  • You choose the output length, handy for fixed slots
  • Prompt-guided delivery
  • Two resolutions, so drafts stay cheap

What it costs

From 169 tokens ($1.69) per clip.

SettingsTokensPrice
~5s169$1.69
~10s325$3.25
~14s480$4.80
~19s634$6.34

Prices come from the live catalog at $0.0100 per token, the pay-as-you-go rate; subscriptions include tokens at a lower rate. The exact price is shown before you generate, and a failed generation is refunded.

Questions

Why does InfiniTalk ask for a length?
Because it generates a fixed number of frames rather than following the track. Choosing the length up front means the clip already fits your slot when it lands.
What if my audio is shorter than the InfiniTalk length I chose?
The run is refused before it costs anything, with a message saying so. Either pick a shorter length or give it a longer track.
Does InfiniTalk need a prompt?
Yes. A short description of what the person is doing is required. The other talking-head models treat it as optional.

Use it inside

Related models