Effective Date: October 15, 2026
This AI Addendum (“Addendum”) supplements your Terms of Service or other agreement (“Terms”) with Arcade Software, Inc. ("Arcade"). By accepting this Addendum, providing Input Data to the Service, or using the Service, you agree to be bound by this Addendum. The AI Features are considered a part of the “Service” under the Terms and all other terms and conditions of the Terms are incorporated by reference and will remain in full force and effect. If any term of this Addendum conflicts with the Terms, this Addendum will control solely with respect to the subject matter it addresses — namely the processing of Input Data by AI Features and rights in Output. This Addendum does not vary the limitation of liability, indemnity, confidentiality, or termination provisions of the Terms, and does not vary any Data Processing Addendum between the parties, except that this Addendum controls with respect to the meaning of Model Training and Service Evaluation and Improvement and Arcade's use of Input Data for those activities, and the allocation of rights in Input Data and Output. Where Arcade and Customer have entered into a separately negotiated written agreement addressing AI Features, that agreement controls over this Addendum.
IF YOU ARE ENTERING INTO THIS ADDENDUM ON BEHALF OF A COMPANY OR OTHER LEGAL ENTITY, YOU REPRESENT THAT YOU HAVE THE LEGAL AUTHORITY TO BIND THE ENTITY TO THESE TERMS, IN WHICH CASE “CUSTOMER” OR “YOU” (WHETHER OR NOT CAPITALIZED) MEANS THE ENTITY YOU REPRESENT. IF YOU DO NOT HAVE SUCH AUTHORITY, OR IF YOU DO NOT AGREE WITH THE ADDENDUM, YOU SHOULD NOT ACCEPT THIS ADDENDUM AND MAY NOT USE THE SERVICE.
AI Features use artificial intelligence to enhance your Arcades, including:
To deliver these features, Input Data, including screenshots and recordings of the application screens you capture, is transmitted to the AI Subprocessors listed in Appendix A, which process it to generate the outputs you request. Section 2 describes how that data is handled.
Arcade may remove or introduce additional AI Features from time to time.
All Arcade Services incorporate AI Features. Certain AI Features are integral to the Service and cannot be operated separately from it; others may be enabled or disabled at the workspace level as described in Section 1.2.
Arcade may enable, modify, or discontinue individual AI Features over time, and configures their availability at the workspace level. If and where an AI Feature can be operated independently of the rest of the Service, Customer may request that Arcade enable or disable it for Customer’s workspace by contacting Customer’s Customer Success Manager or [email protected]. Arcade will use commercially reasonable efforts to accommodate such requests but is not obligated to do so where the request would materially impair the Service or require disproportionate engineering effort.
Customer retains all right, title, and interest in and to all content Customer submits to AI Features, including text, images, brand settings, and other data ("Input Data"). Input Data is considered Content under the Terms. Nothing in this Addendum transfers ownership of Input Data to Arcade.
Arcade and our AI Subprocessors process Input Data solely to provide AI Features and generate the outputs Customer requests (“Output”), except as otherwise described in this Section 2. As between Customer and Arcade, and to the extent permitted by applicable law, Customer owns Input Data and Output. Customer grants Arcade a limited, non-exclusive, worldwide license to use Input Data solely as necessary to (a) provide the AI Features, (b) publish and display Input Data within the AI Features at Customer’s direction, (c) provide support for the AI Features, (d) enforce Arcade’s or its licensors’ policies, (e) prevent fraud, (f) comply with applicable law, and (g) maintain the safety and security of the AI Features, (h) conduct Service Evaluation and Improvement as described in Section 2.3, and (i) provide Workspace Personalization, where enabled for Customer's workspace, as described in Section 2.5. Arcade may not use Input Data for any purpose not expressly permitted in this Section 2.2 or elsewhere in this Section 2. Output may not be unique, and the AI Features may provide the same or similar Output to other users. Customer is solely responsible for determining the intellectual property rights Customer has in Output and its commercial usability, taking into account, among other things, Customer’s usage scenarios and the laws of the relevant jurisdiction.
For purposes of this Addendum, "Model Training" means using Input Data to create, adjust, or influence the weights, parameters, or embeddings of any machine-learning model, including pre-training, fine-tuning, instruction-tuning, or reinforcement learning. "Service Evaluation and Improvement" means using Input Data to test, evaluate, or improve the operation of the Services, including prompts, orchestration logic, configuration, and quality-assurance processes, provided that Input Data (i) is not used for Model Training and (ii) is not incorporated into any prompt, template, or example reused across customers.
Arcade does not use Input Data for Model Training, whether for Arcade's own use or for use by any third party. In particular, Input Data from Customer's workspace will not be used to:
If Arcade later offers a different option to train or fine-tune a model using Customer's data, that feature will be: (a) clearly described at the time of enablement, and (b) made available only on an explicit opt-in basis under separate terms.
Service Evaluation and Improvement is not Model Training and is permitted under this Addendum. It includes routine quality assurance, evaluation, benchmarking, tracing, and abuse monitoring conducted by or on behalf of Arcade that do not modify model weights. Automated safety and abuse-detection monitoring conducted by an AI Subprocessor on its own behalf is addressed in Section 2.4. Arcade limits Input Data used for Service Evaluation and Improvement to what is necessary for that purpose and applies the same confidentiality, access-control, and security measures that apply to all other Input Data.
Arcade contractually requires each AI Subprocessor identified in Appendix A to use privacy-preserving and no-training modes where available. Unless Customer expressly approves otherwise in writing, AI Subprocessors:
Input Data retained under clause (ii) is used solely to detect and prevent violations of that subprocessor's usage policies, is not used for Model Training, and is not made available to Arcade. On Customer's request, Arcade will provide the retention period published by an AI Subprocessor for Input Data processed under this Addendum, or direct Customer to that subprocessor's published policy.
Arcade may store Input Data and Output within the Service to support product functionality for Customer’s workspace (for example, to enable Customer to retrieve and edit previously generated content). Such stored data is not used to train AI models and is retained only for so long as necessary to provide the Service or as required by applicable law.
Where this feature is enabled for Customer’s workspace, Arcade may use Customer's previously published Arcades to inform and personalize Output generated for that same workspace (“Workspace Personalization”). Workspace Personalization operates only within Customer’s workspace, does not share Customer’s data with other customers, and does not modify the parameters of any AI model. Workspace Personalization is configured by Arcade at the workspace level; Customer may request that Arcade enable or disable it for Customer’s workspace at any time by contacting Customer’s Customer Success Manager or [email protected].
Customer may request deletion of Input Data and Output. Upon receipt of a valid deletion request, Arcade will delete such data from its production systems and from its primary third-party storage and media-processing providers within sixty (60) days. That obligation does not extend to (a) copies held in automated backup, archival, or disaster-recovery systems, which will be deleted in the ordinary course of Arcade’s backup rotation cycle (not to exceed ninety (90) days), or (b) usage, telemetry, and analytics records that have been de-identified or aggregated such that they cannot reasonably be used to identify Customer or any individual. Data retained under clause (a) will remain subject to Arcade's confidentiality and security obligations for so long as it is retained.
AI Subprocessors retain Input Data only as described in Section 2.4. Where an AI Subprocessor provides a mechanism to delete Input Data on request, Arcade will submit Customer’s deletion request to that subprocessor promptly after receiving Customer’s request.
Where Arcade and Customer have entered into a Data Processing Addendum, the deletion provisions of that addendum control over this Section.
Input Data retained by an AI Subprocessor under Section 2.4(ii) for automated safety and abuse-detection monitoring is not subject to deletion on request and will expire in accordance with that subprocessor's retention period.
Arcade uses certain third-party infrastructure and model providers (“AI Subprocessors”) to deliver the AI Features described in this Addendum. AI Subprocessors process Input Data solely to generate the requested Output for Customer’s workspace, except as described in Section 2.4. Arcade maintains written agreements with each AI Subprocessor that impose data protection obligations that are, in substance, the same as those imposed on Arcade under this Addendum and, where a Data Processing Addendum is in place between Arcade and Customer, under that addendum.
A complete and current list of AI Subprocessors is provided in Appendix A to this Addendum and is also published at arcade.software/ai-terms, which Arcade will keep current.
Arcade may update Appendix A from time to time as necessary to support the AI Features. Where Arcade and Customer have entered into a Data Processing Addendum, changes to AI Subprocessors are subject to the notice and objection provisions of that addendum, which control over this Section. Where no Data Processing Addendum is in place, Arcade will publish the updated Appendix A at arcade.software/ai-terms before the change becomes effective, and continued use of the AI Features following the effective date constitutes acceptance of the updated list.
If you have questions or need clarification regarding an AI Subprocessor, please contact your Customer Success Manager or [email protected].
Customer is responsible for reviewing all Output before publishing or distributing it. Customer shall conduct a human review where required by applicable law (including, where applicable, Article 22 of the GDPR and analogous provisions under U.S. state privacy laws) and as appropriate given Customer’s use case. Output may:
Customer is responsible for ensuring Output complies with:
Without limiting the disclaimers in the Terms, AI Features and Output are provided "AS IS” and “AS AVAILABLE." Arcade does not warrant that Output will be accurate, complete, error-free, free of infringing content, or suitable for Customer’s specific purposes. Customer acknowledges that AI technology is inherently probabilistic and that Output may contain errors, hallucinations, or biases.
For the avoidance of doubt, the limitation-of-liability provisions in the Terms governing Customer’s use of the Service apply in full to the AI Features and Output. Nothing in this Addendum creates any liability of Arcade to third parties.
Arcade may update this Addendum from time to time. The current version and effective date are available at arcade.software/ai-terms.
For material changes that reduce Customer’s rights or increase Customer’s obligations, Arcade will provide at least thirty (30) days’ advance notice via email to Customer’s designated contact or through the Service. Any such change will not take effect until the start of Customer's next Subscription Term or Billing Cycle. Continued use of the Service after that date constitutes acceptance. Arcade may elect to charge a fee for use of AI Features that are not already included in Customer's Subscription Plan, in accordance with the "Fee Changes" section of the Terms.
If Customer does not agree to a material change to this Addendum, Customer may stop using the affected AI Features within the notice period. Where an AI Feature is integral to the Service and cannot be disabled independently, Customer’s sole and exclusive remedy is to cancel the affected Subscription before the change takes effect, either through Customer's online account management page or by contacting Arcade at [email protected]. To the extent applicable, update procedures follow the amendment terms in the Terms or other governing agreement with Arcade.
Arcade will, upon Customer’s reasonable written request and at Customer’s expense, provide information reasonably necessary for Customer to comply with applicable privacy and AI-related laws and regulations, including without limitation responding to data subject access requests, supporting Customer’s data protection impact assessments, and providing documentation regarding AI Feature functionality and data handling practices. Arcade will notify Customer without undue delay if Arcade becomes aware of any regulatory investigation or enforcement action relating to the AI Features that may materially affect Customer.
The following entities may process Input Data on Arcade's behalf to provide the AI Features, and, where indicated below, for their own automated safety and abuse-detection monitoring or to enable specific product functionality. The listed location identifies the principal processing location used by Arcade; processing may also occur in other locations permitted by the applicable agreement and transfer mechanism.
Retention of Input Data by AI Subprocessors is governed by Section 2.4. Where an entry above describes a specific retention practice, that practice is permitted under Section 2.4 or, where noted, falls outside it.
