AI Disclosure

Every output produced by SeriaLab is AI-generated. We use the following third-party models. The license that applies to your generated output is the license of the model that produced it; review these before commercial use.

Scripts and character bios

Anthropic Claude Sonnet 4.6. Subject to the Anthropic acceptable use policy. Generated text is owned by you to the extent permitted by their commercial terms.

Images

fal.ai — Nano Banana (primary), Flux Pro 1.1 (fallback). Subject to the underlying model licenses. Commercial use generally permitted on paid fal.ai tiers, which SeriaLab uses on your behalf.

Video

fal.ai — Kling 2.1 (short clips), Veo 3 (long clips). Each clip's license is shown in the Episode Editor before generation. Veo 3 outputs carry an invisible SynthID watermark per Google's requirements.

Composition

fal.ai ffmpeg-api. No model is used — output is a deterministic concatenation of the approved scene clips.

Moderation

OpenAI Moderation API classifies prompts and outputs for safety violations. We do not store prompt content with OpenAI beyond the moderation call.

What we promise

• Every paid generation is gated through pre-gen safety checks.
• Generations that fail post-gen safety checks are blocked and refunded.
• We do not train any model on your prompts or outputs and do not let upstream providers do so where opt-out is available.