Praxim

The clause allows indefinite, perpetual, or necessity-based retention.

“2.1 Limited License. Subject to these Terms, Praxim grants you a revocable, non-exclusive, non-transferable, limited license to install and use the add-in and access the Service solely for your internal use. 2.2 Reservation of Rights. Praxim and its licensors retain all rights, title, and interest in and to the Service, including software, models, interfaces, and all intellectual property. No implied licenses. 2.3...”

The clause imposes arbitration, class-action waiver, or jury-trial waiver terms.

“If you opt out, disputes will be resolved in the courts specified in Section 20.1. 16.8 Costs. Payment of filing, administration, and arbitrator fees will be governed by the AAA Rules. To the extent any fees are not allocated by the AAA Rules, the parties will split them equally; provided that if you demonstrate to the arbitrator that you are unable to pay your portion, or the arbitrator determines it is appropria...”

The clause imposes arbitration, class-action waiver, or jury-trial waiver terms.

“20.1 Governing Law. These Terms are governed by the laws of the State of California , without regard to conflicts rules, subject to the arbitration provisions above. For claims not subject to arbitration (and for injunctive relief), the state and federal courts in San Francisco, California have exclusive jurisdiction and venue, and you consent to personal jurisdiction there. 20.2 Entire Agreement. These Terms (and...”

The clause includes sublicensable, transferable, or assignable rights.

“2.1 Limited License. Subject to these Terms, Praxim grants you a revocable, non-exclusive, non-transferable, limited license to install and use the add-in and access the Service solely for your internal use. 2.2 Reservation of Rights. Praxim and its licensors retain all rights, title, and interest in and to the Service, including software, models, interfaces, and all intellectual property. No implied licenses. 2.3...”

The clause grants a broad content license.

“2.1 Limited License. Subject to these Terms, Praxim grants you a revocable, non-exclusive, non-transferable, limited license to install and use the add-in and access the Service solely for your internal use. 2.2 Reservation of Rights. Praxim and its licensors retain all rights, title, and interest in and to the Service, including software, models, interfaces, and all intellectual property. No implied licenses. 2.3...”

model training improved from high/training permitted to low/no training claim.

Latest stance: no training claim on training use

“We use information to: Provide and operate the Services (including editing, vector search, transcription, and related features). Authenticate and secure accounts and sessions; prevent fraud/abuse; detect, investigate, and respond to security incidents. Maintain and improve the Services (quality, reliability, performance, usability, and safety). Measure and manage subscriptions, feature entitlements, usage limits, and billing. Communicate about updates, policy changes, security alerts, and support. Comply with law and enforce our Terms of Service and acceptable use rules. Create aggregated/de-identified insights (which do not identify you) for analytics and service improvement. What we store and don't store: We do not retain prompts or chat history beyond processing as required to provide the Services, unless you explicitly opt in to a feature that requires retention or we need to retain limited data to investigate abuse, comply with law, ensure security, or provide technical support and debugging. We do not use your Inputs, Outputs, or other data to train models . From your inputs to the application, we store uploaded files and user preferences to maintain them across sessions. Other data (prompts, conversations, document content) is generally processed in real time and may be retained as necessary to provide the Services, except as described above or as necessary for technical support, debugging, and service improvement. Important: Third-party AI and service providers may process your data under their own terms and privacy policies, which apply to their services and, where applicable, your use of such features constitutes consent to such processing.”

Latest stance: third party or vendor sharing on training use

“We use information to: Provide and operate the Services (including editing, vector search, transcription, and related features). Authenticate and secure accounts and sessions; prevent fraud/abuse; detect, investigate, and respond to security incidents. Maintain and improve the Services (quality, reliability, performance, usability, and safety). Measure and manage subscriptions, feature entitlements, usage limits, and billing. Communicate about updates, policy changes, security alerts, and support. Comply with law and enforce our Terms of Service and acceptable use rules. Create aggregated/de-identified insights (which do not identify you) for analytics and service improvement. What we store and don't store: We do not retain prompts or chat history beyond processing as required to provide the Services, unless you explicitly opt in to a feature that requires retention or we need to retain limited data to investigate abuse, comply with law, ensure security, or provide technical support and debugging. We do not use your Inputs, Outputs, or other data to train models . From your inputs to the application, we store uploaded files and user preferences to maintain them across sessions. Other data (prompts, conversations, document content) is generally processed in real time and may be retained as necessary to provide the Services, except as described above or as necessary for technical support, debugging, and service improvement. Important: Third-party AI and service providers may process your data under their own terms and privacy policies, which apply to their services and, where applicable, your use of such features constitutes consent to such processing.”

Latest stance: third party or vendor sharing on prompt ownership

“We disclose information to the following categories of recipients (only as needed to provide and protect the Services or comply with law): Cloud hosting and infrastructure providers (e.g., compute, storage, networking). Identity, authentication, and security providers. AI and model providers to process prompts/embeddings/transcriptions and deliver outputs you request. Productivity platform connectors you choose to enable (e.g., file import/export integrations). Payment processors for subscription billing and fraud prevention. Analytics/telemetry processors to understand performance and reliability (where used). Professional advisors (legal, accounting) and potential acquirers in a corporate transaction. Government and law enforcement when legally required or to protect rights, safety, or the integrity of the Services. We do not allow service providers to use personal information for their own independent purposes unrelated to the Services. Key Third-Party Services: We partner with several third-party providers to deliver the Service, including: OpenAI and Microsoft Azure: We use OpenAI and Microsoft Azure services for large language model (LLM) inference to provide AI-powered editing, content generation, and related features. Amazon Web Services (AWS): We use AWS for file storage and authentication/login services. These providers are enterprise-grade services that maintain industry-leading security standards and compliance certifications (such as SOC 2, ISO 27001, and others).”

Latest stance: liability limited on indemnity liability

“The Services are intended for use in the United States , and our Microsoft Word Store listing is geofenced accordingly. If you access the Services by sideloading or otherwise from outside the United States, you are solely responsible for compliance with your local laws. To the extent permitted by law, we disclaim liability for use of the Services outside supported regions.”