| Language: |
Aira / Meeting the World Legal and Content Statement
Effective Date: 2026-09-05
Developer and Publisher: wmdm
Support and Privacy Contact: wmdm314@outlook.com
© 2026 wmdm. All rights reserved for original content; third-party content and its licenses shall prevail where applicable.
I. Ownership
The original program, interface design, branding, and content created or lawfully licensed by the developer belong to their respective rights holders. This product includes third-party software, AI models, Live2D technology, and sample data, which are governed by their respective licenses and terms of use. Use of the product does not transfer any third-party intellectual property rights.
II. Generative Artificial Intelligence
Aira uses AI to generate conversation, optional voice output, and interactive decisions. The action AI selects actions from existing action templates and does not generate new Live2D models. This Windows version includes the Gemma 4 e2b local chat model; users may also configure third-party AI services. CosyVoice3 synthesizes speech locally using built‑in or user‑selected reference recordings. Icons and some localization content were produced with the assistance of generative AI.
AI outputs may be inaccurate, incomplete, or inappropriate, and do not represent the views of the developer. They must not be relied upon as a basis for high‑risk decisions in medical, legal, financial, or emergency contexts. The program imposes restrictions and validations on tool execution, action selection, and resource files, but these do not constitute semantic review of all chat content, nor can they guarantee that every output is lawful, safe, or non‑infringing. This statement does not replace the safety measures that the developer should implement.
When using external APIs, selected character cards, necessary conversation context, enabled emotion and cognition information may be sent to the chosen service provider. When autonomous action is enabled, unsent chat drafts may also be used for action determination or proactive replies; when screen tools are enabled, the model may capture screen images when invoking such tools. This is not a pre‑screenshot for every chat, nor does it guarantee that each screenshot is separately prompted each time. Specific processing methods are detailed in the accompanying privacy policy and the service providers’ policies.
III. User Content and Steam Workshop
This fixed‑avatar version uses only the built‑in Hiyori and disables local Live2D model import, replacement, Workshop model upload and download, and online model file transfer. Voices, character cards, and their preview images may still be imported, created, selected, or published according to the features provided by the product.
Users are responsible for ensuring that they have the necessary copyright, voice usage authorisation, privacy consent, and other permissions for importing, processing, and publicly sharing content. Merely being able to lawfully listen to a recording does not grant the right to clone, redistribute, or publish the voice therein. Users must not upload infringing content, malicious files, credentials, or unauthorised personal data.
Importing from a private library or selecting such content does not constitute public distribution. When a user actively publishes a voice to Steam Workshop, its reference recording may be made available to other users along with the work; users should confirm before publishing that they are willing to make it public and have the right to distribute that recording. Publishing, subscribing, downloading, and deleting are also subject to the applicable Steam agreement and Workshop terms. The developer does not claim ownership over user‑generated original content.
IV. Data and Privacy
Conversations, character cards, cognition data, settings, and managed assets are generally stored locally, but external AI, web tools, Steam networking, and Workshop will process data necessary to complete their respective functions; “local storage” does not mean that these functions are fully offline.
When long‑term memory is enabled, relevant cognition data may be saved across conversations and restarts, and sent to enabled external AI characters for use; in local mode, existing cognition data is retained but frozen. Unsent drafts are used in‑process and are not saved as sent user messages, nor does the behaviour log record the draft text; retention rules of external service providers for received requests apply separately.
API keys are stored via the operating system’s secure storage mechanism; diagnostic exports attempt to filter sensitive fields, but users should still review them before sharing. The current version only provides local error logging and user‑selected local diagnostic exports; it does not automatically upload error reports.
Clearing history and factory reset are not the same function: the former retains user‑imported assets, while the latter may clear Aira‑managed data and imported copies, but does not delete original files outside the management directory. Deleting local files does not automatically delete copies already stored by third‑party service providers, Steam, or other users. The viewing method in the settings interface does not affect the privacy rights users enjoy under law; relevant requests may be sent to the email above.
V. Live2D Statement
This product uses Live2D Cubism technology and the fixed built‑in Hiyori sample model. Live2D, Cubism, and related names and logos are owned by Live2D Inc. The sample data is provided under the applicable free material license and sample model terms, and does not grant users any right to redistribute or commercially use them.
This content uses sample data owned and copyrighted by Live2D Inc.
This product is independently created by wmdm and has no operational partnership or endorsement relationship with Live2D Inc. Third‑party data is subject to the original terms accompanying it.
VI. Third‑Party Software and Models
This product involves Electron/Chromium, PixiJS, Ollama, Gemma 4, CosyVoice, VTubeStudioJS, Steamworks, and other components. Different components and their sub‑components are subject respectively to the MIT, BSD, Apache 2.0, LGPL, MPL, Live2D proprietary terms, Steamworks agreement, and other licenses; a single MIT or Apache label cannot cover the entire package.
The accompanying “Third‑Party Notices”, `licenses`, and component list describe the copyright and licensing of each component. For the respective components, the third‑party licenses prevail; this statement does not restrict the rights to inspect, modify, replace, relink, redistribute, or reverse engineer for debugging modifications as permitted by such licenses. The corresponding source code and means of obtaining it must be provided in accordance with the relevant licenses; the license text list itself does not substitute for source code delivery obligations. Third‑party names are used for identification only and do not imply endorsement, sponsorship, or partnership.
VII. Network Functions
Steam networking, matchmaking, Workshop, external APIs, and web functions are subject to service status, network and regional conditions, and successful connection is not guaranteed at all times. In AI conversation mode, Aira’s replies, voice, and actions are exchanged; in human relay mode, messages actively sent by the user along with corresponding translations and optional AI comments are transmitted. The product does not transmit the user’s microphone audio through these two modes, and the fixed‑model version does not transmit model files.
VIII. Limitation of Liability and Statutory Rights
To the extent permitted by applicable law, this product is provided “as is” and does not warrant that AI outputs are accurate, that voice is fully similar, or that third‑party services are uninterrupted. The scope of liability is governed by applicable law; this statement does not exclude or limit any liability of the developer that cannot be lawfully excluded, nor does it exempt the developer from its obligations regarding the product and published content.
Nothing in this statement excludes or limits statutory rights that consumers cannot waive.
IX. Applicable Documents and Contact
Please read this statement, the accompanying End User License Agreement, and the Privacy Policy before use. When using Steam or third‑party services, their applicable agreements also take effect; third‑party licenses prevail for the corresponding components. This statement does not designate an exclusive jurisdiction and does not exclude the application of mandatory laws.
For support, privacy, and rights‑related matters, please contact wmdm314@outlook.com.
Effective Date: 2026-09-05
Developer and Publisher: wmdm
Support and Privacy Contact: wmdm314@outlook.com
© 2026 wmdm. All rights reserved for original content; third-party content and its licenses shall prevail where applicable.
I. Ownership
The original program, interface design, branding, and content created or lawfully licensed by the developer belong to their respective rights holders. This product includes third-party software, AI models, Live2D technology, and sample data, which are governed by their respective licenses and terms of use. Use of the product does not transfer any third-party intellectual property rights.
II. Generative Artificial Intelligence
Aira uses AI to generate conversation, optional voice output, and interactive decisions. The action AI selects actions from existing action templates and does not generate new Live2D models. This Windows version includes the Gemma 4 e2b local chat model; users may also configure third-party AI services. CosyVoice3 synthesizes speech locally using built‑in or user‑selected reference recordings. Icons and some localization content were produced with the assistance of generative AI.
AI outputs may be inaccurate, incomplete, or inappropriate, and do not represent the views of the developer. They must not be relied upon as a basis for high‑risk decisions in medical, legal, financial, or emergency contexts. The program imposes restrictions and validations on tool execution, action selection, and resource files, but these do not constitute semantic review of all chat content, nor can they guarantee that every output is lawful, safe, or non‑infringing. This statement does not replace the safety measures that the developer should implement.
When using external APIs, selected character cards, necessary conversation context, enabled emotion and cognition information may be sent to the chosen service provider. When autonomous action is enabled, unsent chat drafts may also be used for action determination or proactive replies; when screen tools are enabled, the model may capture screen images when invoking such tools. This is not a pre‑screenshot for every chat, nor does it guarantee that each screenshot is separately prompted each time. Specific processing methods are detailed in the accompanying privacy policy and the service providers’ policies.
III. User Content and Steam Workshop
This fixed‑avatar version uses only the built‑in Hiyori and disables local Live2D model import, replacement, Workshop model upload and download, and online model file transfer. Voices, character cards, and their preview images may still be imported, created, selected, or published according to the features provided by the product.
Users are responsible for ensuring that they have the necessary copyright, voice usage authorisation, privacy consent, and other permissions for importing, processing, and publicly sharing content. Merely being able to lawfully listen to a recording does not grant the right to clone, redistribute, or publish the voice therein. Users must not upload infringing content, malicious files, credentials, or unauthorised personal data.
Importing from a private library or selecting such content does not constitute public distribution. When a user actively publishes a voice to Steam Workshop, its reference recording may be made available to other users along with the work; users should confirm before publishing that they are willing to make it public and have the right to distribute that recording. Publishing, subscribing, downloading, and deleting are also subject to the applicable Steam agreement and Workshop terms. The developer does not claim ownership over user‑generated original content.
IV. Data and Privacy
Conversations, character cards, cognition data, settings, and managed assets are generally stored locally, but external AI, web tools, Steam networking, and Workshop will process data necessary to complete their respective functions; “local storage” does not mean that these functions are fully offline.
When long‑term memory is enabled, relevant cognition data may be saved across conversations and restarts, and sent to enabled external AI characters for use; in local mode, existing cognition data is retained but frozen. Unsent drafts are used in‑process and are not saved as sent user messages, nor does the behaviour log record the draft text; retention rules of external service providers for received requests apply separately.
API keys are stored via the operating system’s secure storage mechanism; diagnostic exports attempt to filter sensitive fields, but users should still review them before sharing. The current version only provides local error logging and user‑selected local diagnostic exports; it does not automatically upload error reports.
Clearing history and factory reset are not the same function: the former retains user‑imported assets, while the latter may clear Aira‑managed data and imported copies, but does not delete original files outside the management directory. Deleting local files does not automatically delete copies already stored by third‑party service providers, Steam, or other users. The viewing method in the settings interface does not affect the privacy rights users enjoy under law; relevant requests may be sent to the email above.
V. Live2D Statement
This product uses Live2D Cubism technology and the fixed built‑in Hiyori sample model. Live2D, Cubism, and related names and logos are owned by Live2D Inc. The sample data is provided under the applicable free material license and sample model terms, and does not grant users any right to redistribute or commercially use them.
This content uses sample data owned and copyrighted by Live2D Inc.
This product is independently created by wmdm and has no operational partnership or endorsement relationship with Live2D Inc. Third‑party data is subject to the original terms accompanying it.
VI. Third‑Party Software and Models
This product involves Electron/Chromium, PixiJS, Ollama, Gemma 4, CosyVoice, VTubeStudioJS, Steamworks, and other components. Different components and their sub‑components are subject respectively to the MIT, BSD, Apache 2.0, LGPL, MPL, Live2D proprietary terms, Steamworks agreement, and other licenses; a single MIT or Apache label cannot cover the entire package.
The accompanying “Third‑Party Notices”, `licenses`, and component list describe the copyright and licensing of each component. For the respective components, the third‑party licenses prevail; this statement does not restrict the rights to inspect, modify, replace, relink, redistribute, or reverse engineer for debugging modifications as permitted by such licenses. The corresponding source code and means of obtaining it must be provided in accordance with the relevant licenses; the license text list itself does not substitute for source code delivery obligations. Third‑party names are used for identification only and do not imply endorsement, sponsorship, or partnership.
VII. Network Functions
Steam networking, matchmaking, Workshop, external APIs, and web functions are subject to service status, network and regional conditions, and successful connection is not guaranteed at all times. In AI conversation mode, Aira’s replies, voice, and actions are exchanged; in human relay mode, messages actively sent by the user along with corresponding translations and optional AI comments are transmitted. The product does not transmit the user’s microphone audio through these two modes, and the fixed‑model version does not transmit model files.
VIII. Limitation of Liability and Statutory Rights
To the extent permitted by applicable law, this product is provided “as is” and does not warrant that AI outputs are accurate, that voice is fully similar, or that third‑party services are uninterrupted. The scope of liability is governed by applicable law; this statement does not exclude or limit any liability of the developer that cannot be lawfully excluded, nor does it exempt the developer from its obligations regarding the product and published content.
Nothing in this statement excludes or limits statutory rights that consumers cannot waive.
IX. Applicable Documents and Contact
Please read this statement, the accompanying End User License Agreement, and the Privacy Policy before use. When using Steam or third‑party services, their applicable agreements also take effect; third‑party licenses prevail for the corresponding components. This statement does not designate an exclusive jurisdiction and does not exclude the application of mandatory laws.
For support, privacy, and rights‑related matters, please contact wmdm314@outlook.com.