TAU-HOME.COM
LOADING

AI Wednesday Clip in 4 Steps: From Image Generation to Dance Swap

A 4-step AI virtual-character clip workflow shared by X user @kelmruns: image generation, stage selection, Seedance 2.5 swap, and hashtag-only captions, plus ve

tau · October 10, 2026

#AIcharacter #Seedance #image-generation

AI Wednesday Clip in 4 Steps: From Image Generation to Dance Swap

The original author is X user @kelmruns (kelm), who shared a 4-step AI virtual-character clip recipe in a post dated 2026-10-10. The figures of 67.4M views and $27,433 from a single clip are the author's claims; replies in the same thread dispute the payout structure and the model attribution.

Pale girl with two black braids and blunt bangs dancing in a plain white room, AI virtual character clip concept

Image source: Image: @kelmruns on X

This article preserves the post's procedure in its original order and adds the verification points raised in the replies. It does not independently confirm the revenue or model claims.

The 4 steps as posted

The procedure below keeps the original wording and order.

  • Generate a pale girl with two black braids and blunt bangs, styled like Wednesday from the Netflix series, in ChatGPT.
  • Pick a plain white room and the dance everyone already knows.
  • Swap her into the dance with Seedance 2.5.
  • Caption it with hashtags only.

Verification points from the replies

Three issues were raised in the reply evidence.

  • Revenue figure: the original claims $27,433 from one clip. A reply (@renorivsan) disputes this based on short-form payout structure. With platform, period, and payout method unconfirmed, treat it as a claim only.
  • Copyright and likeness risk: replies (@restapai, @schlampvibes) flag that monetizing a character resembling a Netflix series character and actor likeness can create rights problems. Review rights before commercial use.
  • Model attribution: the original says it was made "with GPT Astra 6." A reply (@flowersslop) notes there is no evidence the image step used that model. Quote the model name as written without stating it as verified fact.

Practical notes

  • The 4-step recipe assumes a consistent character look; the original gives no separate reference-locking method.
  • The white room is the author's staging choice; the original does not state an edge-concealment effect.
  • Hashtag-only captions are the author's style; reach effects are unverified.

Original source