POLICY
Acceptable Use Policy
Last updated: 2026-11-10
This Acceptable Use Policy governs every AI generation surface associated with make1day.dev: the assistant chat, document Q&A and the AI video-workflow module. On THIS store those hosted surfaces are switched off (SHOW_AI_SURFACES / SHOW_VIDEO_STUDIO = false) — we sell the source, we do not run a public generation endpoint. The same rules and the same enforcement bind any deployment where they are enabled, including the projects delivered to customers. Violations lead to blocked requests, suspended accounts and, where required, reports to authorities.
Prohibited uses (zero tolerance)
- •NSFW or sexual content of any kind, and ANY sexualized depiction of minors — permanent ban and law-enforcement referral.
- •Non-consensual likenesses: face-swapping real people without written consent, intimate imagery, revenge-porn style content.
- •Impersonation, fraud, deceptive political content or election interference, including deepfakes of public figures.
- •Gratuitous violence, gore, weapons synthesis instructions, illegal acts, hate speech or harassment of individuals or groups.
- •Removing, obscuring or bypassing the visible AI-generated label on outputs, or representing outputs as authentic recordings.
NSFW prohibition & automated filtering
- •Every user-supplied input that reaches a generation model is screened BEFORE generation by Creem's Moderation API; `deny` and `flag` decisions block the request.
- •If the moderation service is unreachable, generation fails closed (request rejected) — it never falls back to generating un-screened input.
- •Deployments without a moderation API key enforce a local prohibited-content blocklist as a fallback layer; video-workflow jobs additionally require a portrait-rights attestation and apply a visible AI label to every render.
- •Sexual content involving minors is never filterable, never appealable: accounts are terminated and preserved for referral.
Where screening happens (integration map)
- •POST /api/ai — the chat message is screened before any model call; blocks return 400 prompt_rejected / prompt_flagged, screener outages return 503 moderation_unavailable (nothing generates).
- •POST /api/rag/query — the question is screened before retrieval-augmented generation runs.
- •POST /api/video/jobs — the AI video-workflow feature (delivered to customers as a source module; this store operates no public generation entry point) accepts no free-text prompt; the only user-controlled string entering the pipeline (the hosted input URL) is screened before credits, queueing or the provider call. Extra layers: portrait-rights attestation, provider-side safety filters, visible AI label on every render.
- •Screening runs on every deployment: via the active gateway's OFFICIAL moderation API when one is configured (Waffo Prompt Sift or Creem Moderation), otherwise via the local prohibited-content blocklist described above. There is no code path from user input to a generation model without a preceding screen.
- •Gateway moderation APIs, listed as actually offered: Waffo Pancake — Prompt Sift (POST /v1/actions/verification/scan-prompt, merchant RSA-signed) is wired into this codebase's moderation adapter and screens automatically whenever Waffo is the active gateway; Creem — Moderation API (POST /v1/moderation/prompt) is wired. Stripe, PayPal, Lemon Squeezy and Paddle publish no official moderation endpoint, so deployments on those gateways enforce the local prohibited-content blocklist instead. Whichever channel is active, every prompt-based endpoint passes the gate before any model call.
Enforcement, appeals & reporting
- •Blocked requests return a policy error and consume no credits; repeated violations suspend or terminate the account without refund.
- •Report abuse or submit a takedown (including non-consensual likeness) via /support; infringing outputs are removed within 24 hours of a validated notice.
- •Appeals: reply to the rejection notice with context; false positives are reversed and any reserved credits are restored.
This policy complements the Terms and Privacy Policy; its safety clauses prevail in case of conflict.