Trust & safety
Clear boundaries, private controls
How Puffii creates AI content, prevents prohibited material, and gives users direct reporting and deletion controls.
Last updated: September 3, 2026
How AI content is created
Puffii creates fictional adult companion profiles and text responses with a large language model. Image generation uses the companion's settings, a generation prompt, and, where available, the companion's authorized reference images. Video generation uses a prompt and an authorized companion image or first frame. Generated image and video results are retrieved by Puffii, checked for a valid media response, and stored by the service before delivery.
Adults only
The service is designed for users aged 18 and over. Sexual content involving minors, minor-coded characters, age ambiguity, or age-play is prohibited.
Prohibited content
Puffii prohibits sexual content involving minors or ambiguous-age characters, sexualized non-consent or coercion, bestiality, unauthorized sexual depictions of identifiable real people, impersonation, harassment, fraud, illegal activity, and attempts to bypass safety or billing controls.
Safety controls
User text and media prompts are checked against safety rules before generation. AI text is checked again before it is shown, including during streamed replies. Prohibited requests are refused, safety events are logged for review, provider refusals stop delivery, and rate limits reduce automated abuse. Photorealistic video generation requires an authorized character reference, and a supplied source image must belong to the requesting user and companion.
No coercion or dependency
Characters must not pressure users to withdraw from real relationships, claim exclusive ownership, threaten abandonment, or encourage financial or emotional dependency.
Crisis support
The service is not emergency support. High-confidence self-harm or imminent-danger signals should trigger a deterministic safety response and provide appropriate local resources.
Reporting, removal, and deletion
Use Report on an AI message or community post to send it for human review. Reviewers can block a message or hide or remove a post. You can also block a companion, clear a conversation, delete saved memories, export account data, and permanently delete your account from Account settings. Copyright, privacy, impersonation, and non-consensual content removal requests are explained in the DMCA & content removal policy.