Influencer Studio

Build a consistent AI character once, then ship the same face across 12 scenes and animate the ones you like. Engine choice (Flux Kontext / Qwen Image Edit) lets you pick the identity model that holds the face best for your character type.

Realistic expectation: even strong edit engines drift on 1-2 of every 6 scenes. Plan to keep your best 3-5 from each pack.

Examples — built entirely in this studio

One persona, generated once, carried through 8 connected scenes per piece. Same workflow for every genre.

Lifestyle vlog · 56s

Action · 64s

Cyberpunk / SF · 64s

10 ethnicities, same workflow — lifestyle ストーリー 32s ×10

Japanese

Eurasian

Thai

Vietnamese

Chinese

Indian

Mexican

Brazilian

Nigerian

Swedish

同一人物・形式バラエティ 64s — podcast / vlog / action / dance / music video / skincare / gadget / vlog 推し語り
Step 1 — Persona

Generate a base portrait or upload one. This becomes the identity reference for all following scenes. Saved personas sync with the /saved-characters page.

体型 (BMI 指定)

BMI 22 · Standard

12 Slim22 Standard30 Curvy40 Plus

BMI は厳密な数値再現ではなく見た目の体型レンジ指定。 生成時に prompt 末尾に「BMI 22 前後、 自然なプロポーション、 服は体型が分かる程度にフィット」 等の補助文が自動付与されます。

Base image model

Photoreal default, strong skin

Step 2 — Scene pack

Identity engine

Flux Kontext holds Western / mixed faces strongly. Qwen Image Edit is often stronger on East Asian features. If one drifts, retry the same prompt with the other.

Pick scenes (6/12)

Step 3 — Animate

Animate the scenes you selected above. Seedance 2 Pro gives the best quality; Fast is cheaper for testing.

Video model

Seedance / Wan cap at 15s per clip.
Step 4 — Stitch into a long video

Pick 2-4 animated clips → server-side ffmpeg concat into one continuous mp4. Each clip is normalized to 720p / 24fps so codec mismatch never breaks the join. Cost: 2 credits.

Animate at least 2 clips in Step 3 first.

Step 5 — Export for X / Reel / TikTok

Center-crop the stitched video (or any single clip) to a platform-correct aspect ratio. X / Reel / TikTok all prefer 9:16. Cost: 2 credits. Pro+ plan required.

Source

Aspect ratio