FussAI FREE

FUSSAI FREE – CREATIVE AI SERVICE TERMS AND NON-COMMERCIAL LICENCE Version 2026-09-v1.21 Effective: 28 September 2026 1. PROVIDER AND SERVICE SCOPE FussAI is provided by Jiří Holý, Company ID 06230181, registered office Uralská 689/7, 160 00 Prague 6 - Bubeneč, Czech Republic, e-mail admin@lajkto.eu, operating the FussyCraft / Lajkto / FussAI projects. FussAI is a creative generative AI service. Depending on functions currently released, it may include AI chat, image analysis, image generation and editing, video generation or editing, audio and music generation and related future creative AI functions. Not every function must be available at the same time. 2. FREE TIER AND NON-COMMERCIAL USE An output created under the FREE tier is intended for personal and other non-commercial use unless a specific function expressly states otherwise. Purchasing membership later does not retroactively convert a FREE output into a commercially licensed PRO output. The Provider grants only such rights to an output as the Provider is legally able to grant. The Service does not guarantee that a purely AI-generated output is protected by copyright, is exclusive, or will not be similar to an output generated for another user. 3. FUSSAI AI CREDITS AND PUBLIC PRICING FussAI may use a FREE allowance, membership AI credits or FussAI AI credits for individual actions. The credit cost of an action is governed by the public pricing page and the amount displayed in the Service before use. A FussAI AI credit is not legal tender or electronic money. FussAI AI credits are a general accounting unit for FussAI creative generative services and are not limited to music. 4. RIGHTS IN INPUTS – MANDATORY USER WARRANTY A user may submit or make available to FussAI text, photographs, illustrations, video, audio, music, recordings, voices, a person's likeness, logos, brands, documents or other content only if the user has all rights, licences, permissions and consents required for the intended AI processing and intended use of the result. It is prohibited to submit third-party photographs, music, recordings or video merely because they are publicly accessible online. Public availability by itself is not permission for AI processing. Users must not submit song lyrics written by another person, including translated or adapted versions, unless they are authorised to use them in the service. The user is responsible in particular for copyright, performers' and phonogram/videogram producers' rights, trade marks, trade secrets, rights to likeness and voice, privacy and personal data, and for obtaining consent from identifiable persons where legally required. 5. PEOPLE, VOICE, LIKENESS AND DEEPFAKES FussAI must not be used for unauthorised impersonation, false endorsement, fraud, extortion, harassment, reputational harm, deception of the public or other unlawful interference with the rights of an identifiable person. Without an appropriate legal basis or required consent, the user must not create or distribute a realistic manipulation of another person's likeness or voice in a manner capable of falsely appearing to be an authentic recording of that person's act or statement. Where an output is an image, audio or video deepfake or other AI-generated or manipulated content, the user as the person deploying or publishing that output must comply with all applicable transparency duties, in particular Article 50 of Regulation (EU) 2024/1689 (AI Act), including disclosure of artificial generation or manipulation where required by law. For evidently artistic, creative, satirical or fictional works, the disclosure scope provided by law applies. The user must not remove or circumvent machine-readable or visible AI-origin markings where such marking is required by law or forms part of a Service safety measure. 6. PROHIBITED USE FussAI must not be used for unlawful content or conduct, including threats, targeted harassment, fraud, phishing, identity theft, unlawful discrimination, hateful incitement or violent threats directed at persons or groups, extremist or terrorist propaganda contrary to law, unlawful sexual content, sexual abuse or sexualisation of minors, non-consensual intimate content, unlawful disclosure of private information, infringement of intellectual-property rights or other unlawful conduct. Creative, documentary, educational, journalistic or satirical context is assessed according to the specific purpose and applicable law and does not authorise an unlawful interference with third-party rights. Without the Provider's prior express written permission, the user must not knowingly upload, submit, transmit, make available or otherwise provide a FussAI output, in whole or in substantial part, as an input, reference, training example or data source to any generative or other artificial-intelligence system operated by a third party. This prohibition includes, in particular, use of the output for training, fine-tuning, distillation, dataset creation, embeddings or reference sets, voice or style cloning or imitation, derivative generation, remixing, continuation, rearrangement, conversion, separation, restoration, mastering, or other analysis or transformation performed by third-party AI. Ordinary publication or distribution of an output through a music, video, social or other platform is not a breach of this prohibition where the user does not intentionally submit the output to a third-party AI feature and cannot reasonably control the platform's internal automated, recommendation, safety or moderation systems. The prohibition also does not apply to AI functions expressly provided by the Provider within FussAI. A knowing or repeated breach of this prohibition may be treated as a serious breach of these Terms. 7. MODERATION, WARNINGS AND ACCOUNT DEACTIVATION Where there are reasonable grounds to suspect a breach of these Terms or law, the Provider may refuse a prompt or generation, restrict a feature, hide or disable access to an output, warn the user, temporarily suspend the account, or deactivate the account for serious or repeated breaches. In cases of immediately serious abuse, restriction may occur without a prior warning where necessary to protect persons, the Service or comply with a legal duty. Knowingly or repeatedly submitting unauthorised third-party lyrics or other protected content may be treated as a serious or repeated breach. The Provider may retain necessary security and evidentiary records to the extent permitted by law and cooperate with competent authorities where legally required or permitted. Moderation does not transfer responsibility for the user's own use of an output to the Provider. 8. USER RESPONSIBILITY AND PROVIDER POSITION The user is responsible for prompts, submitted material, model instructions, selection and editing of outputs, publication, distribution, disclosure and subsequent use. FussAI does not authorise infringement of third-party rights, and an output is not a legal clearance of the user's intended use. The Provider is not the author of the user's prompt or the approver of the purpose for which the user later uses an output. To the maximum extent permitted by law, the Provider is not responsible for damage or claims arising from unlawful or contractually prohibited use of the Service attributable to the user. Nothing in these Terms excludes or limits rights or liability that cannot lawfully be excluded, including liability relating to natural personality rights, intent or gross negligence to the extent provided by Section 2898 of the Czech Civil Code, and mandatory consumer rights. 9. GENERATIVE NATURE AND OUTPUT REVIEW AI may make factual, visual, audio or other errors, hallucinate, generate material similar to other content, or fail to follow a prompt precisely. The user must reasonably review an output before important or public use. The Provider does not guarantee absolute originality, exclusivity, legal clearance, commercial success, acceptance by a distribution/social platform or a specific artistic result. 10. TECHNICAL AVAILABILITY AND REFUNDS Functions may be limited by queues, compute capacity, maintenance or external infrastructure. A paid action that technically fails before usable delivery is not charged or the applicable consumed credit is restored according to the Service rules. Subjective dissatisfaction with a technically delivered creative result does not by itself create an automatic entitlement to another credit; mandatory defect/non-conformity rights remain unaffected. 11. ORDINARY INPUTS, PUBLIC LIBRARY AND IMPROVING FUSSAI Ordinary prompts and files used for a specific chat, analysis or generation are processed to provide the Service, operate it securely and perform necessary technical processing. Uploading content, creating an output or storing it privately in the Library is not by itself consent to model training or fine-tuning. If the user voluntarily makes a music output public and actively confirms its placement in the Public library, the user grants the Provider a non-exclusive, royalty-free, worldwide and purpose-limited permission to technically process and analyse that specific public music output, reproduce it to the extent necessary for computation, convert technical formats, segment it, measure audio and statistical properties, and use it to select quality examples, evaluate, test, train, fine-tune or otherwise improve current and future FussAI models and systems. The Provider may select only a subset of public outputs, especially outputs marked by an authorised reviewer as TOP quality. A private prompt, private text source or other private input is not automatically included in this permission. The user also confirms that they are allowed to give the Provider this limited permission for the public output. The purpose is not to take ownership of the particular track, present it as the Provider's own work, or separately sell or distribute it. The purpose is to teach the generative system general technical and quality characteristics of successful outputs, such as clean and natural vocals, guitar, bass, drums and other instrument quality, spatial character, dynamics, rhythm, arrangement, timbre, mix, mastering and overall audio stability, so that future generation can hold a comparable quality level more reliably. The user retains rights in the user's own human creative contribution. This limited training permission does not transfer ownership of the track or give the Provider a right to a share of the user's ordinary distribution or monetisation. The non-monetary benefit of voluntary sharing is collective service improvement: high-quality public outputs help the models learn what users regard as successful sound, so future generations can progressively improve for the contributor and for other users. Public sharing by itself does not create a right to monetary compensation, revenue share or Contributor Units; any such compensation is governed only by the separate Contributor Program or another express agreement. If the user removes the output from the Public library or makes it private, it will not be newly included in future training runs after that point where reasonably technically and legally feasible. Removal does not automatically have retroactive effect on models or adapters that were lawfully trained earlier, unless mandatory law requires otherwise. Outputs made public before this revision takes effect are not treated as authorised for new training use without a new active confirmation by the user. 12. PRIVACY AND PERSONAL DATA Personal-data processing is governed by the Lajkto/FussAI Privacy Policy and Regulation (EU) 2016/679 (GDPR) where applicable. Users must not submit third-party personal or sensitive information without a lawful basis. 13. CONSUMERS AND DIGITAL CONTENT Where the user is a consumer, statutory consumer rights apply. Where the Service requires an express request for immediate supply of digital content or service before expiry of a withdrawal period and the corresponding statutory acknowledgement, that consent is recorded separately. Nothing in these Terms limits rights that a consumer cannot waive. 14. LEGAL FRAMEWORK AND ILLEGAL-CONTENT NOTICES These Terms are governed by Czech law and directly applicable European Union law. Relevant rules include Act No. 89/2012 Coll., the Czech Civil Code, Act No. 121/2000 Coll., the Copyright Act, Act No. 634/1992 Coll., the Consumer Protection Act, Regulation (EU) 2024/1689 (AI Act), Regulation (EU) 2022/2065 (Digital Services Act) where applicable, and the GDPR. Notices of potentially unlawful content or abuse may be sent to admin@lajkto.eu. The current general Lajkto Terms, Privacy Policy and moderation rules also apply unless these special Terms govern a matter in greater detail. 15. EVIDENCE OF ACCEPTANCE AND CHANGES On electronic acceptance the system may record account identity, terms version, SHA-256 hash of the exact text, date/time, IP address, user-agent and active acknowledgements. A historically accepted text is not rewritten retroactively. A material new revision applies to future use after the form of notice or renewed acceptance required by law and the Service. 16. OUTPUT RIGHTS AND HUMAN CONTRIBUTION The Provider does not claim authorship or ownership of the user's prompt, original lyrics or other user input. Rights in human-created elements remain with their lawful rights holders. The Service does not guarantee that a purely or predominantly AI-generated output qualifies for copyright protection, is exclusive, or gives the user “copyright” in the entire output. FREE mode remains contractually limited to non-commercial use under these Terms. If the Provider acquires a transferable right in a specific FREE output, the user receives only the permission corresponding to the non-commercial scope of FREE mode unless the relevant feature expressly states otherwise. 17. PLATFORMS, AI DISCLOSURE AND CONTENT ID The user must make any AI disclosure or labelling required by applicable law, a distributor or a platform. FussAI does not guarantee acceptance by a particular platform or eligibility for YouTube Content ID or any other automated rights-management system. A generative model may unintentionally produce an output that resembles an existing work or recording in style, melody, harmonic progression, rhythm, structure, arrangement, vocal phrasing, instrumentation, tempo, key, or a combination of those elements. A single common musical element, such as a key, tempo or ordinary chord progression, does not by itself necessarily indicate infringement; the particular combination of elements, degree of similarity and overall musical impression may be relevant. Before publicly releasing, distributing or monetising a music output, including through Spotify, Apple Music, YouTube, YouTube Music, TikTok, a distributor or a similar platform, the user must reasonably listen to and review the output for originality and for any conspicuous similarity to a specific existing protected work or recording. If the user identifies or reasonably suspects substantial similarity, the user must not distribute that output without resolving the necessary rights and should edit it, regenerate it or obtain the required permission. FussAI cannot compare every output against every existing song and recording in the world and does not warrant that an output will be free from similarity to third-party content or that downstream use will not infringe third-party rights. Similarity may arise unintentionally from the generative nature of the model, common musical patterns, the user prompt or chance, and similarity by itself does not establish that the model retrieved a specific song from training data or deliberately copied it. 18. THIRD-PARTY MODEL, IMAGE COVER AND CREATIVE ML OPENRAIL-M The optional cover generator may use third-party components subject to applicable use restrictions. Users must comply with those restrictions, including prohibitions on unlawful use, harm to minors, harmful disinformation, harmful use of personal data, defamation and harassment, discriminatory or harmful automated decision-making, exploitation of vulnerabilities, medical advice or interpretation of medical results, and predictive profiling for justice, law-enforcement, immigration or asylum processes. The applicable restrictions form part of the Service rules. A generated cover is a separate AI image output produced from available inputs and metadata. It may not factually, semantically, narratively, visually or stylistically match the lyrics, audio, title, mood or subject of the track. The Provider does not guarantee that a cover matches the track content or is suitable for a particular distribution, marketing or artistic purpose. The user must review the cover before publication or distribution. 19. FUSSAI 1.0, FREE LIMITS AND MUSIC GENERATION Under this revision, FussAI offers a single current product version for music generation: FussAI 1.0. The FussAI 1.5 and FussAI 2.0 product branches are not offered. FREE mode allows no more than 10 music outputs per calendar month. Each accepted FREE request consumes one unit of the monthly FREE allowance. A FREE music output may be no longer than 180 seconds and is processed at a fixed 8 inference steps. Where the Provider's local RTX 5060 Ti compute capacity is immediately available, the request may be processed locally. If it is unavailable, the Provider may use backup cloud capacity for the same FussAI 1.0 product, preferring RTX 3090 and, where unavailable, RTX 4090. This operational choice does not change the product version or the user price. A technically unfinished request is not treated as a duly delivered output. If the applicable unit or amount has already been deducted, the refund rules in these Terms apply. Future changes to published limits or prices apply only to new actions after publication and do not apply retroactively. 20. SOUND AI – EFFECTS, DUBBING AND TRANSCRIPTION FussAI may provide separate audio features, including text-to-sound-effect generation, speech synthesis and dubbing, optional synthesis using a reference voice (voice clone), and speech-to-text transcription. These are separate creative AI actions and are not music generation. The current production prices under this revision are: a sound effect up to 15 seconds costs 1 AI credit; a sound effect longer than 15 seconds and up to 30 seconds costs 1.5 AI credits; standard dubbing/TTS up to 30 seconds costs 1.5 AI credits; dubbing with a reference voice up to 30 seconds costs 2 AI credits; transcription costs 0.5 AI credit for each commenced minute of processed input. The binding price for a particular action is always displayed before it starts; future price changes apply only prospectively. If the compute capacity required for a sound effect is unavailable, the action may be refused, queued within the scope shown by the Service, or terminated technically. A more expensive paid variant will not start without the user's explicit confirmation. A paid audio action that fails technically before delivery is not charged or the amount actually deducted is refunded. Any processing time shown in the interface, including a target or estimate such as “within 30 seconds”, is an operational target or estimate rather than a guaranteed SLA unless expressly agreed in writing. 21. REFERENCE VOICE, VOICE CLONE AND IDENTITY A user who uploads or uses a reference voice expressly represents that it is their own voice or that they have sufficient authorisation and consent from the person concerned for recording, upload, AI processing, synthesis, imitation and the intended use of the result. This representation also covers any applicable performer rights, personality rights, privacy and personal-data rights. Voice clone or dubbing must not be used for fraud, impersonation, false endorsement, bypassing voice authentication, obtaining access to an account, social engineering, extortion, misleading the public or any other unlawful interference with another person's rights. Technical acceptance of a reference recording by FussAI does not confirm that the user has the necessary rights. 22. AUDIO CONTENT RETENTION AND AUTOMATIC DELETION FussAI is not a long-term archive or backup service. User content created or uploaded in Sound AI, including reference voice recordings, other audio inputs, sound effects, dubbing outputs and text transcripts, may be retained in active storage for no longer than 365 days from the creation or upload of the relevant item. Once that period is reached, the system may automatically and irreversibly delete the oldest items without further notice. Content may also be deleted earlier at the user's request, for security reasons, due to a breach of the Terms, technical failure, or reasonable capacity management, provided mandatory user rights are not infringed. The user is responsible for timely downloading and backing up any content they wish to keep long term. Expiry of the retention period or deletion after an output was duly made available does not create a right to a credit refund. The 365-day limit applies to user audio content itself and its normal operational copy. It does not apply to separate accounting, tax, contractual, transaction and security records that do not contain the audio content itself and that the Provider may or must retain for a different period under applicable law. After content deletion, minimal technical metadata or a cryptographic hash may be retained where reasonably necessary to evidence a transaction or refund, prevent abuse, or comply with a legal obligation. 23. AVAILABILITY AND DATA LOSS Unless expressly agreed otherwise, the Provider does not guarantee uninterrupted availability, a particular processing time, continued availability of a particular model, preservation of history or recovery of a deleted file. The user is responsible for backing up duly delivered outputs they intend to continue using. Mandatory consumer rights and liability that may not be excluded or limited by law remain unaffected.