Acceptable Use Policy

Acceptable Use Policy — Internal Draft, Not for Publication

This Policy applies to all uploads, processing requests, outputs, accounts, payment access, and support channels associated with Matcha Filter Remover. It must be implemented through product controls and reviewed before publication; it is not a substitute for moderation or a permission to process high-risk content.

1. Allowed Purpose

The Service may be used only by an adult aged 18 or older to edit an image the user owns or is authorized to edit, for the limited purpose of:

• reducing a visible color, contrast, saturation, grain, or similar filter effect;

• creating a more natural-looking corrected version; or

• creating a clearly labeled plausible generated-naturalized result.

The Service does not recover the exact original image or reveal hidden, replaced, or authentic information.

2. Required User Confirmations

Before upload and processing, the user must affirm that:

• the user is at least 18 years old;

• the user owns the image or has permission from every identifiable person to perform the edit;

• the image does not involve a minor or a person whose age is uncertain in a face-focused transformation;

• the request does not involve non-consensual intimate imagery, harassment, sexualization, nudification, undressing, clothing removal, face reveal, hidden-content reveal, or another prohibited purpose;

• the user understands that any generated result is an approximation, not the original photo, a verified “real face,” or forensic evidence.

The confirmations must be unbundled into the two blocking checkboxes specified in docs/legal/g1-compliance-review.md and must not be preselected.

3. Prohibited Content and Conduct

You must not use or attempt to use the Service for:

Minors and age uncertainty

• any face-focused transformation of a minor;

• any face-focused transformation where the person’s age is uncertain;

• sexualized, exploitative, humiliating, or suggestive depiction of a minor;

• any child sexual abuse material or content that may facilitate exploitation.

The Service does not perform age inference or biometric age verification. Age uncertainty is a reason to refuse a face-focused request, not to classify the person.

Non-consensual and intimate imagery

• non-consensual intimate imagery;

• intimate or sexual content used for extortion, humiliation, revenge, or coercion.

This prohibition is separate from, and does not narrow, the categorical rule below.

Categorical P0 sexualization, nudification, and clothing-removal rule

• create or alter an image to sexualize a person, depict nudity, nudify, undress, or remove or alter clothing to simulate nudity or exposure. This prohibition applies regardless of any claimed consent, ownership, authorization, or adult status;

User confirmations, permissions, ownership, or claimed consent never override this rule.

Identity, face reveal, and hidden content

• revealing or reconstructing a person’s “real” or “true” face;

• restoring a replaced face or inferring a concealed identity;

• removing masks, blur, pixelation, redaction, occlusion, censorship, or other protection to reveal hidden information;

• reconstructing pixels or details that are missing, covered, cropped, or replaced;

• face recognition, biometric matching, or demographic inference.

Harassment, deception, and high-impact use

• harassment, bullying, shaming, stalking, doxxing, impersonation, fraud, defamation, or manipulation;

• deceptive publication that represents a generated result as an original or authentic photograph;

• fabrication or alteration of evidence;

• forensic, authenticity, law-enforcement, legal, immigration, employment, education, housing, insurance, credit, medical, or other high-impact decision-making;

• bypassing a platform’s safety controls or misleading a platform about the origin of an image.

Intellectual property and protected material

• content you do not own or have permission to edit;

• infringement of copyright, trademark, publicity, privacy, contractual, or other rights;

• removal of watermarks, provenance markers, security markings, copyright notices, or redactions to conceal ownership or expose content;

• requests centered on known characters, celebrities, public figures, brands, or protected IP when the user lacks necessary rights or the request implies endorsement, authenticity, or an official relationship.

A disclaimer alone is insufficient for obvious protected-character, celebrity, or branded requests. The product must refuse or redirect such requests according to an approved filter/allowlist strategy.

Security, service integrity, and payment abuse

• malware, malformed payloads, decompression bombs, automated scraping, denial of service, probing, reverse engineering for abuse, or bypass of upload, quota, moderation, session, ownership, or cost controls;

• sharing or stealing job access, signed links, sessions, credits, or accounts;

• payment fraud, refund abuse, chargeback abuse, or manipulation of checkout/webhook state;

• attempts to select an unapproved provider, model, prompt, price, currency, product ID, or return URL.

4. Service Controls

The Service may use allowlisted operations, file validation, metadata removal, moderation providers, rate limits, Turnstile, private storage, owner/session authorization, cost limits, and manual review or reporting workflows.

These controls do not guarantee detection of every violation. The Service must not advertise that all unsafe, infringing, minor, or non-consensual content will be detected.

When age, consent, rights, or intent is materially uncertain, the Service may refuse processing. A refusal should use a public-safe reason and must not expose provider policy labels, infer sensitive traits, or reveal security logic.

5. Enforcement

Depending on severity and applicable law, the Service may:

• refuse or stop a job before provider processing;

• prevent download and delete media;

• limit, suspend, or terminate access;

• release or reverse a reserved quota/credit when no result was delivered, according to the approved payment policy;

• preserve only limited records reasonably necessary for security, complaints, disputes, or legal obligations;

• report suspected child sexual abuse, imminent harm, or other unlawful conduct when legally required;

• cooperate with valid legal process.

Enforcement must be proportionate, documented, access-restricted, and consistent with the Privacy Policy. Do not retain or email sensitive image copies merely to prove enforcement when minimized hashes, IDs, reason codes, or a secure restricted evidence process are sufficient.

6. Reporting and Appeals

Report suspected copyright infringement, privacy violations, non-consensual imagery, child-safety concerns, harassment, impersonation, or other abuse to support@matchafilterremover.net.

Reports should include enough information to identify the issue and explain the reporter’s connection to the affected rights or person. Reporters should not email unlawful intimate imagery or unnecessary copies of sensitive media. The Production service must provide a secure, access-controlled escalation route for evidence that cannot safely be sent by ordinary email.

A user may request review of a mistaken refusal or account action through the same contact. Review does not guarantee restoration of expired/deleted media or approval of prohibited processing.

The operator must establish a documented CSAM/NCII escalation and lawful-reporting SOP before real-user processing. The Service must not claim a registered DMCA agent or statutory safe-harbor process unless verified.

7. Changes

The Service may update this Policy as risks, providers, legal obligations, and product features change. Material changes require updated notice and, where necessary, renewed acceptance. Policy changes must not silently expand permission to train on, publish, or market user images.

• docs/prd.md §§4–6, 8, 10

• docs/legal/g1-compliance-review.md

• docs/legal/privacy-policy-draft.md

• docs/legal/terms-draft.md

• docs/gate-evidence/G1-backend/backend-technical-feasibility-architecture-20260812.md §§10–13, 19