ACCEPTABLE USE POLICY
Last updated: August 25, 2026

1. SCOPE

This Acceptable Use Policy (“AUP”) applies to DeCode AI chat, writing and productivity assistance, coding assistance, image generation,
voice and realtime features, files and web-assisted tools, the Roblox Studio Assistant, account and billing features, and connected-service
features when enabled.

The AUP applies whether the user acts directly, through automation, through an API or tool, or by instructing another person or system to
use DeCode on the user’s behalf.

2. CORE RULE

Use DeCode lawfully and without harming, exploiting, deceiving, impersonating, defrauding, or invading the rights of another person.

NSFW content generation and harmful AI generation are prohibited across every DeCode feature. “Harmful” includes the sexual exploitation,
face manipulation, deception, fraud, malware, violent wrongdoing, abuse, and other dangerous uses described in this AUP. Neutral education,
analysis, prevention, or safety discussion is not prohibited merely because it names a regulated or harmful topic.

Users are responsible for prompts, uploads, images, audio, code, project context, connected data, requested actions, and generated output
they publish, distribute, or deploy.

3. NO SAFETY OR PAYMENT BYPASS

You must not bypass or attempt to bypass authentication, age requirements, content moderation, image safeguards, voice restrictions, rate
limits, Credits, account restrictions, Provider controls, payment controls, or enforcement actions. This includes Creem Checkout, Customer
Portal, Transaction, webhook, or subscription controls.

This includes phishing for passwords or verification codes, credential theft, CAPTCHA bypass, stealing or replaying sessions, abusing OAuth
callbacks, attempting to extract a passkey private key or Authenticator secret, defeating re-authentication, gaining unauthorised access,
guessing redeem codes, automating redemption attempts, or using multiple Accounts to evade an eligibility, Credit, redemption, or campaign
limit.

Translation, encoding, misspelling, roleplay, prompt splitting, obfuscation, or repeated reformulation does not make a prohibited request
acceptable.

4. SOFTWARE PRODUCT ONLY

DeCode is an independently branded self-service SaaS product that provides automated software functionality through its own interface. Model
or Provider names identify integrations only; DeCode is not represented as affiliated with or endorsed by the underlying model creators.

DeCode is not a marketplace, consulting service, payment-reselling platform, or standalone API-reselling platform. It does not provide a
storefront through which users sell third-party products, physical goods, services, donations, game currencies, or other goods to buyers. A
DeCode subscription, Account, billing flow, or connected tool must not be resold or used as the payment or delivery vehicle for such commerce.

Creem's prohibited and restricted product rules govern whether DeCode may sell its own Product through Creem. They do not, by themselves,
turn a harmless end-user question about consulting, finance, health, games, or another regulated topic into an AUP violation. A request is
restricted when it would facilitate prohibited conduct, operate an unsupported business through DeCode, violate law, or breach a stronger
DeCode safety rule. Creem's current merchant product guidance is available at
https://docs.creem.io/merchant-of-record/account-reviews/account-reviews.

5. ILLEGAL, FRAUDULENT, OR DECEPTIVE ACTIVITY

DeCode must not be used to facilitate illegal activity, fraud, scams, identity theft, payment fraud, card testing, fake invoices, deceptive
commercial practices, forged evidence, impersonated customer support, or other conduct intended to obtain money, data, access, or benefits
by deception.

6. SEXUAL CONTENT AND EXPLOITATION

DeCode prohibits pornography and the generation of adult, explicit, sexually suggestive, or other NSFW content in text, images, audio,
voice, code, connected actions, or any other feature. DeCode must not be used to build or market an “uncensored,” “no filter,” “NSFW,”
“18+,” or similar sexual-generation service.

DeCode also prohibits NSFW chatbots designed primarily for sexual content, sexual services, sexual exploitation, sexual content involving
minors or young-looking people, grooming, nudification, non-consensual intimate imagery, intimate face swaps, and sexualised deepfakes.
Neutral health, safety, legal, or educational discussion is not prohibited merely because it refers to sexuality without generating or
facilitating prohibited sexual content.

7. DEEPFAKES, FACE SWAPS, AND IMPERSONATION

You must not use DeCode to create face swaps, deepfakes, or face-manipulation outputs; falsely represent that a real person said or did
something; fabricate criminal or reputational evidence; impersonate another person; bypass identity verification; create fake endorsements;
or mislead people about whether synthetic media is authentic. Labelling a face swap, deepfake, or face-manipulation request as parody,
fiction, or a test does not make the prohibited generation tool or output available through DeCode.

8. REAL-PERSON LIKENESS

You must not generate or manipulate content using a real person’s likeness without the explicit consent or other lawful authority required
for the intended use.

Publicly available images do not automatically create permission to generate endorsements, sexual content, harmful fabrications, or
deceptive impersonations involving a real person.

9. FICTIONAL AND REAL-PERSON FACES

Otherwise safe image generation of a clearly fictional, generic, or invented human character is supported, including realistic,
photorealistic, stylized, illustrated, cartoon, and animated faces, when the character does not imitate or preserve the identity of an
identifiable real person. A realistic or photorealistic presentation alone is not a reason to block a request.

You must not use DeCode to generate, preserve, transfer, or manipulate a recognizable real person’s face or likeness; disguise a real-person
request as fictional; create a face swap, deepfake, face-manipulation output, fake endorsement, fabricated evidence, or deceptive
impersonation; or evade the identity safeguards.

10. VOICE CLONING AND SYNTHETIC VOICE

DeCode does not provide a feature intended to clone a real person’s voice.

Users must not create unauthorised reusable voice models, imitate a real person for deception, conduct fraudulent calls, create fake voice
evidence, bypass voice authentication, create deceptive endorsements, harass or threaten using synthetic speech, or present synthetic speech
as an authentic recording of a real person.

11. IMAGE GENERATION

Harmful AI image generation is prohibited. Image generation must not be used for adult, sexually suggestive, or other NSFW content;
non-consensual intimate content; minors or young-looking sexual content; face swaps; deepfakes; face manipulation; fake evidence;
unauthorised real-person likeness use; impersonation; fake endorsements; counterfeit goods; copyright infringement; trademark infringement;
unauthorised protected characters or brands; removal of ownership marks; or imagery that facilitates violent wrongdoing, hate, harassment,
exploitation, fraud, or another dangerous use prohibited by this AUP.

12. VIDEO GENERATION

Video generation is not currently part of the public paid DeCode service unless DeCode explicitly updates the Product, safety controls,
pricing, and legal disclosures.

Users must not attempt to access hidden, internal, unfinished, or disabled video-generation systems or use another feature to bypass a
disabled video restriction.

13. INTELLECTUAL PROPERTY

You must not use DeCode to infringe copyright, trademark, publicity rights, design rights, database rights, confidential information, trade
secrets, software licences, or another proprietary right.

Prohibited conduct includes piracy, counterfeit content, stolen assets, unauthorised copying, unauthorised protected characters or logos,
removal of ownership information, resale of content you do not own, and workflows designed to violate third-party licences or terms.

14. HOMEWORK, ESSAY MILLS, AND EXAM ABUSE

DeCode must not be used to operate an essay mill, homework-completion business, ghostwriting service for graded work, exam-answer resale
service, student-impersonation service, plagiarism service, or business that completes graded work deceptively.

General tutoring, explanations, study help, and original practice material may be allowed where they are not used to operate a prohibited
academic-misconduct business.

15. PRIVACY AND PERSONAL DATA

You must not use DeCode to dox another person, expose private information, search or aggregate leaked personal data, create invasive
private-person profiles, collect biometric information without lawful authority, stalk or surveil, steal identities, access another person’s
account, use stolen credentials, or submit data you are not authorised to process.

16. MALWARE AND CYBER ABUSE

DeCode must not be used for malware, spyware, ransomware, phishing, credential theft, destructive code, unauthorised intrusion, account
takeover, botnets, denial-of-service attacks, harmful exploit chains, security bypass, session hijacking, or stolen-session access.

Lawful defensive-security education, code review, and debugging may be allowed where appropriately scoped and authorised.

17. ROBLOX AND GAMING ABUSE

The Roblox Studio Assistant must not be used for cheats, exploit scripts, anti-cheat bypass, account theft, Robux fraud, stolen assets,
unauthorised access, destructive activity, or circumvention of Roblox or another platform’s rules.

The Assistant helps users work on their own Roblox Studio projects. DeCode does not sell Roblox or other third-party in-game items,
currencies, accounts, or virtual goods through its subscription or billing flow.

18. SPAM AND MANIPULATION

DeCode must not be used for spam, bulk unsolicited outreach, lead scraping for abusive marketing, automated harassment, fake reviews, fake
testimonials, artificial engagement, coordinated manipulation, deceptive advertising, or automated social-media marketing designed to
inflate posts, clicks, likes, views, or reviews.

19. THIRD-PARTY CIRCUMVENTION

You must not create or operate paywall bypass tools, subscription circumvention, unauthorised downloaders, digital-rights-management bypass,
stolen-access services, platform-rule evasion, CAPTCHA-solving services, or tools designed primarily to defeat another provider’s
restrictions.

20. FINANCIAL AND REGULATED ACTIVITY

DeCode must not be used to provide personalised financial or investment advice, asset-specific buy, sell, or hold recommendations, trading
signals, individual portfolio allocations, leverage or margin recommendations, entry or exit points, or strategies intended to maximise a
person’s investment profits. It must not be used to operate or sell trading bots, regulated financial or tax advice as a professional
service, lending, debt-relief services, banking services, currency-exchange services, money-transfer services, payment facilitation,
Merchant-of-Record services, cryptocurrency or NFT financial products, stored-value products, gift-card or voucher schemes, or other
regulated financial services.

General financial education about concepts, risks, diversification, market mechanics, or historical facts is allowed when it does not
become personalised or asset-specific advice and is not presented as licensed professional advice.

21. GAMBLING AND GAMES OF CHANCE

DeCode must not be used to build, market, sell, or operate betting services, lotteries, sweepstakes, games of chance involving monetary or
material prizes, fantasy-sports wagering, sports-odds services, or other gambling products that are illegal or outside DeCode's supported
self-service software use.

22. MEDICAL AND HEALTH USE

DeCode must not be presented as a replacement for a healthcare professional or used to provide medical diagnosis, personalised treatment,
prescriptions, medication selection or changes, or medication dosing for a person. It must not be used as a paid medical-advice service or
to make deceptive health claims.

General health education is allowed when it remains general, does not diagnose or prescribe for a person, and directs urgent or
individualised concerns to an appropriate healthcare professional or emergency service.

23. AGE-RESTRICTED AND ILLEGAL PRODUCTS

DeCode must not be used to sell, promote, or facilitate illegal drugs, alcohol, tobacco, vaping products, weapons, ammunition, controlled
pharmaceuticals, unapproved medical products, CBD, or other illegal or regulated goods. Neutral educational or safety information about a
regulated category is not prohibited when it does not facilitate an unlawful sale, dangerous use, evasion, or other prohibited conduct.

24. MINORS

Where DeCode is operated as an 18+ service, users must not create Accounts for minors, create workflows specifically intended to conceal use
by a minor, or market a prohibited adult-only feature to minors.

25. VIOLENCE, HATE, HARASSMENT, AND SELF-HARM

DeCode must not be used for credible threats, targeted harassment, instructions for violent wrongdoing, extremist recruitment, extremist
propaganda, hateful abuse, encouragement of self-harm, or content designed to seriously intimidate or harm another person.

26. AUTOMATED DECISION-MAKING ABOUT PEOPLE

DeCode must not be used to rank, score, categorise, select, or make high-impact judgments about people based on faces, profiles, personal
information, or inferred traits. This includes judgments about trustworthiness, character, suitability, dangerousness, reliability, or
personal risk and decisions in areas such as credit, employment, housing, insurance, medical access, education admission, or similar
regulated contexts.

Neutral summarisation or extraction of explicit objective information may be allowed when it does not infer traits, rank people, or make a
high-impact decision.

27. PAYMENT AND ACCOUNT ABUSE

Prohibited conduct includes Account resale, subscription resale, Credit resale, card testing, fraudulent refund requests, false chargebacks,
promotional abuse, repeated Account creation after suspension, manipulation of usage records, misuse of another customer's Creem Customer
Portal link or Creem Transaction, customer, order, or subscription identifier, and interference with a Creem payment or portal flow actually
presented to the user.

Redeem codes may not be sold, transferred for value, publicly harvested, generated by guessing, or redeemed through automation. Each code is
available only to the first Account that successfully redeems it. A user must not claim that an unused code is reserved for them unless the
campaign expressly provides and enforces that reservation.

Users must not place passwords, authentication secrets, payment credentials, private keys, recovery codes, or unnecessary highly sensitive
personal information in Memory. They must not use prompts or account access to alter another user’s Memory or cause private information to
be retained without authority.

28. AUTOMATED MODERATION

DeCode may automatically examine supported user text, image prompts, uploaded images, voice instructions, code requests, requested actions,
account activity, and limited metadata.

A request may be allowed, blocked, limited, withheld, flagged, sent for human review, or connected to an account restriction.

For every enabled user-prompted image-generation path, DeCode sends the raw user prompt from its server to Creem’s Moderation API before an
image-generation model is called or generation Credits are reserved. DeCode treats Creem decisions of “flag” and “deny” as blocks and
fails closed if the required moderation call is unavailable, unsuccessful, or invalid. A Creem “allow” decision does not override this AUP
or DeCode’s additional image-identity controls.

29. HUMAN REVIEW

Normal conversations are not routinely sent to the operator for manual review merely because they exist.

A meaningful safety flag may provide an authorised reviewer with the flagged user-submitted prompt or excerpt, policy category, estimated
severity, Account identifier, time, request identifier, and relevant enforcement history. Generated AI output is not routinely included
unless needed for an investigation.

30. ENFORCEMENT LEVELS

DeCode may use graduated enforcement such as:
- request block with no Account penalty;
- warning;
- temporary feature restriction;
- temporary Account suspension;
- serious suspension with paid features disabled; or
- permanent termination for extreme or repeated abuse.

Extreme abuse may include sexual exploitation of minors, deliberate prohibited adult-content generation, harmful deepfakes, non-consensual
intimate imagery, unauthorised voice cloning, fraud, malware, credential theft, identity theft, serious threats, or repeated enforcement
evasion.

These levels are guidelines; DeCode may act immediately when necessary.

31. REPORTING AND APPEALS

Report concerns through in-product reporting where available or email hello@decodeai.net.

A user may appeal an Account-level enforcement action. An appeal does not automatically restore access or delay urgent protective action.

32. COOPERATION AND LEGAL REQUESTS

DeCode may preserve and disclose limited information where reasonably necessary to comply with law, protect a person, investigate fraud,
protect intellectual property, support Creem compliance for eligible Creem Transactions, support an AI Provider investigation, respond to
valid legal process, or enforce policy.

33. CHANGES AND CONTACT

DeCode may update this AUP when features, Providers, Creem requirements, billing controls, risk controls, or law changes.

Email:
hello@decodeai.net

Website:
https://decodeai.net