n8n-nodes-vaaya-hotel-lobby

n8n community node for the Vaaya Hotel Lobby recipe:
put yourself — or you and a friend — into the viral Hotel Lobby performance
video. Generation runs on MiniMax H3 (minimax/h3/reference-to-video) via
fal.ai, billed to your own fal API key.

Source: github.com/MaruPelkar/n8n-nodes-vaaya-hotel-lobby

Install

In n8n: Settings → Community Nodes → Install and enter
n8n-nodes-vaaya-hotel-lobby.

Credentials

Create a fal.ai API credential with an API key from
fal.ai → Dashboard → Keys. All generations are
billed by fal to that key (H3 is priced per second of video, up to ~$4.68 for
15s at 2K — see fal's pricing page for current rates).

Usage

Node: Vaaya Hotel Lobby.

• Variant — Standard (the full 7-second clip) or Extended (the 29-second
  clip as reference; MiniMax H3 caps output at 15 seconds).
• Left / Right performer photo URLs — 1–5 public image URLs per person, as
  seen from the viewer's perspective in the source video. Fill either slot
  or both; an empty slot keeps the original performer. One clear face photo is
  enough; an outfit shot improves the match.
• Creative notes — optional identity/wardrobe notes. The scene and
  choreography are fixed by the recipe.
• Consent — you must confirm every person shown is you or an adult who gave
  permission. The node refuses to run without it.
• Options — resolution (default 2K), duration override (5–15s), aspect
  ratio (default 4:3, matching the source clip), a custom template video URL,
  and max wait time.

Output per item: video_url (the finished MP4 on fal's CDN), request_id,
and the raw fal response.

Notes

• The reference clips are the same prepared assets the Vaaya product uses,
  served from vaaya.ai/recipe-media/colors/.
• H3 re-generates the scene from references; unlike the hosted Vaaya recipe it
  does not stream-copy the original audio track, so the soundtrack is model-
  generated. For the exact original audio, use the hosted recipe at
  vaaya.ai/colors.
• MiniMax's own content safety checks still apply upstream.

License

MIT