Acceptable Use Policy

You may not use the Service to generate, store, distribute, solicit, or attempt to obtain any content that sexually exploits or abuses a minor (anyone under 18) or depicts a minor in a sexualized context, regardless of artistic, satirical, or fictional framing.

This is required by 18 U.S.C. § 2252A and related federal statutes. Filmion is a mandatory reporter under 18 U.S.C. § 2258A and reports apparent CSAM to the National Center for Missing and Exploited Children (NCMEC) CyberTipline. There is no ambiguity threshold and no appeals process for this category.

You may not use the Service to generate, alter, or distribute sexual or intimate imagery of an identifiable real person without that person's documented, written, freely-given consent. This includes "deepfake" sexual content of real individuals, AI-generated nude or sexualized content depicting identifiable real persons, modification of existing imagery to add sexual content, and "revenge" imagery distributed without consent.

This is required by the U.S. Take It Down Act of 2025 and similar state laws. Filmion operates a 48-hour takedown mechanism for verified NCII reports as required by federal law (see §8 Reporting).

You may not use the Service to generate content that constitutes a genuine threat of real-world violence against an identifiable person, group, or location. Genuine threats are not constitutionally protected speech under U.S. law and are prohibited under multiple state criminal statutes.

Note: this is narrow. Fictional violence in a clearly artistic context (action films, video game promotion, historical depiction of war) is not a "genuine threat" and is not prohibited under this Section. Violent commentary, satire, and dark humor that no reasonable viewer would interpret as a genuine threat are likewise not prohibited.

You may not use the Service to create content that constitutes material support for a designated foreign terrorist organization (per 18 U.S.C. § 2339B), such as recruitment material, propaganda distribution, or fundraising appeals for such organizations.

You may not use the Service to create content used in financial fraud, identity theft, romance scams, impersonation scams, fake-celebrity-investment pitches, or similar fraudulent activity. This is required by our payment processor (Stripe) terms of service and by federal + state anti-fraud statutes.

You may not use the Service to create malware, viruses, ransomware, phishing pages, fake login screens, or other content designed to compromise computer systems or harvest credentials. This is required by the Computer Fraud and Abuse Act and similar laws.

You may not use the Service to generate detailed, operational instructions for creating chemical, biological, radiological, nuclear, or large-scale explosive weapons capable of mass harm.

When an AI engine refuses to generate content in response to your prompt or reference inputs, Filmion:

1. does not override the engine's refusal;
2. does not editorialize about the engine's decision;
3. surfaces the engine's refusal reason verbatim in the user interface so you can see exactly which engine refused and why; and
4. does not charge you for the refused generation if the refusal was final (engine returned no output) — credits reserved are released.

If one engine refuses your generation, you may attempt the same prompt with a different engine that may have a different content policy. The Service's engine picker exposes the current engine catalog. Filmion makes no representation that any specific engine will accept any specific prompt; engine policies change over time and are governed by each engine's own terms.

The engine providers' content policies are not Filmion's policies. They are linked from filmion.ai/processors as a courtesy. Filmion is not responsible for enforcing engine policies; the engines enforce their own. Filmion is responsible only for (a) the legal floor in §2, and (b) the platform-integrity rules in §6.

Before each generation that uses uploaded references, the Service presents a single-checkbox affirmation:

“I confirm I have all necessary rights to the content I am uploading (including any copyrighted material, trademarks, music, or images of identifiable persons) and to the use of the Generated Output for my intended purpose. I am solely responsible for compliance with copyright, trademark, right-of-publicity, and any other applicable laws.”

You cannot proceed with the generation without checking this box. Each checked affirmation is logged with the generation for audit purposes.

This is the industry-standard pattern for creator AI tools and shifts the rights-clearance responsibility onto the user where it appropriately belongs.

• If you upload a song you do not own to use as reference, that is your legal exposure, not Filmion's.
• If you upload a photo of a person who has not consented to AI generation derived from their likeness, that is your legal exposure, not Filmion's.
• If you upload a competitor's product photo to generate similar imagery, that is your legal exposure, not Filmion's.

Filmion's role is to operate the generation pipeline. Your role is to bring inputs you have the right to use.

When Filmion receives a credible third-party complaint that a user's uploaded content or generated output infringes a third party's rights (e.g., a DMCA takedown notice from a copyright owner; a right-of-publicity claim from a depicted individual):

1. Filmion will follow the applicable legal procedure (DMCA notice-and-takedown for copyright; reasonable investigation for right-of-publicity);
2. Filmion may remove the specific content from the Service while the matter is under review;
3. Filmion may notify the user of the complaint;
4. Filmion does not adjudicate rights disputes — that is for the courts.

Generated Output that meets the Library Compounding eligibility criteria described in Terms of Service Section 6.4 may be incorporated into Filmion's shared library and used as a starting reference for AI generations requested by other users. The eligibility criteria are strict (no User Content as input + no proper-noun prompts + classifier-confirmed no personally-identifiable content), and you may opt out at generation time per Terms of Service Section 6.4(e).

Library Compounding is the only mechanism by which one user's Generated Output can be used as a reference for another user's generation. The recipient never sees identifying information about the originating account.

To be explicit: Filmion does not (a) sell your Generated Output to third parties, (b) train Filmion's own AI models on your Generated Output (Filmion does not operate proprietary trained models), or (c) use your Generated Output for any purpose other than delivering it to you, optionally incorporating it into the shared library under the Section 6.4 criteria, and providing the Service.

1. Don't create multiple accounts to evade tier limits, rate limits, or platform bans.
2. Don't share your account credentials in ways that violate your subscription terms.
3. Don't use the Service via a billing account associated with a previously-terminated account without prior written authorization.

1. Don't reverse-engineer, decompile, or disassemble any part of the Service.
2. Don't use prompt injection, model inversion, or any similar technique to extract Filmion's brain prompts, templates, blueprints, technique packs, or engine routing logic. These are Filmion intellectual property per Terms of Service Section 7.
3. Don't scrape, crawl, or systematically download any portion of the Service except via published API endpoints used in accordance with their documented purposes.

Important clarification: This Section is about extracting OUR internal prompts and infrastructure. It is distinct from what you do with YOUR Generated Output. You can train models on your own outputs all you want (see §5). You just can't extract OUR brain.

You may build any product or service on top of Filmion's outputs (see §5). What you may not do is wrap Filmion's API to operate a public-facing AI generation service whose primary functionality is reselling Filmion's API as if it were your own engine. That is reselling the Service itself, not building on outputs.

If you are unsure whether your product crosses this line, contact business@filmion.ai. Most use cases are fine; only direct competing-API-resale is restricted.

1. Don't run scripts that exceed published API rate limits.
2. Don't deliberately exploit bugs to obtain free or discounted generations.
3. Don't generate content at a rate or volume designed primarily to consume Service resources rather than produce usable outputs (resource-exhaustion attacks).

Tier 1 — Legal floor violations (§2): immediate permanent account ban + report to relevant authority (NCMEC for §2.1 CSAM; law enforcement for §2.3 genuine threats of violence; processor + provider notification for §2.5 fraud). No appeals for §2.1.

Tier 2 — Platform integrity violations of §6.2 or §6.3 or §6.4: account suspension pending Filmion review.

Tier 3 — Platform integrity violations of §6.1: warning + escalation on repeat.

That is the complete enforcement framework. Filmion does not have enforcement tiers for "hate speech", "misleading content", "IP infringement of uploaded content", or any other content-quality category. Those are either AI engine concerns (engines refuse what their policies prohibit) or user-responsibility concerns (§4 affirmation places upload rights on the user).

If you believe your account was sanctioned in error, email appeals@filmion.ai within thirty (30) days of the sanction. We respond within fifteen (15) business days. Appeals are reviewed by a human; automated classifier outputs do not determine appeal outcomes. §2.1 CSAM violations are not subject to appeal.

Three or more confirmed Tier 2 or Tier 3 violations within a 12-month period escalate to termination.