Cross-session identity lock

Cross-session identity lock means the spokesperson on Monday is still the spokesperson on Thursday. Same face, same wardrobe rules, same voice performance. One pretty clip is easy. A series is the product.

Three locks, not one toggle

1. Visual identity — reference stills / I2V / Seedance-class reference stacks. Wan’s cross-session identity marketing matters only if you actually reuse the same asset kit.
2. Voice identity — a clone or a designed voice you do not reshuffle weekly.
3. Voice style — style locking in AI TTS so pace and emotion do not drift.

Miss any one and the series feels recast.

A practical kit

• 3–8 stills: front, three-quarter, wardrobe variants you allow
• One voice preset with style lock notes (energy, pace, ban list)
• A lighting recipe in one sentence
• A continuity do-not-change list (hair length, logo side, glasses)

Store them outside the chat so every episode starts from the kit, not from memory.

Episode workflow

1. Generate or select the body with the visual kit (text to video with continuity).
2. Write the VO; synthesize with the locked voice.
3. Lipsync only if the mouth is visible.
4. Burn captions; export; label AI where required.

For UGC ads with gesture variety, keep the still and change the driver — motion control for UGC ads.

FAQ

Is Seedance enough without voice lock? For silent product loops, yes. For a talking series, no.

Can I change wardrobe weekly? Yes if the kit includes those variants. Surprise wardrobe is how lock fails.

Does identity lock bypass YouTube inauthentic rules? No. A consistent face can still be a template farm. Editorial insight still required.

Build the kit once. Sign up and reuse it every episode.