Acceptance of Japan's Generative AI Principles-Code
BYVE AI is a free app for Android and iOS that turns a written description into a phone app that works the way you described.
By My Side Labs ("we") supports the purpose of the Principles-Code for the Protection of Intellectual Property and Transparency for the Appropriate Use of Generative AI adopted by Japan's Intellectual Property Strategy Headquarters on August 25, 2026 (the "Code"), and accepts Principles 1 through 3 of the Code.
Under Section 1(4) of the Code, we publish below how we implement each principle and, where we do not, why.
- Service
- BYVE AI (Android and iOS apps), which generates and edits HTML-based apps that run on the user's device from a written description.
- Review
- We review this page at least once a year and record the date and content of any update here.
Principle 1, disclosure: transparency
Models we use
We do not develop AI models. We use the following models provided by third-party developers through their APIs (as of October 26, 2026).
| Order | Model (identifier) | Provider |
|---|---|---|
| 1 | gpt-6-luna | OpenAI, L.L.C. (United States) |
| 2 | deepseek-flash | Hangzhou DeepSeek Artificial Intelligence Co., Ltd. (China) |
| 3 | qwen3.8-flash | Alibaba Cloud Model Studio (international endpoint) |
| 4 | glm-5.3-flash | Z.ai |
- We mainly use model 1 to generate and edit apps and to suggest features. Models 2 to 4 are fallbacks used automatically when the previous provider cannot respond. Depending on the app version or the user's country (for example, where OpenAI does not offer its API), processing may start with model 2.
- Each model's release dates, revision history and architecture are as published by its developer. We use each model under its developer's API terms, and users need no special hardware or software.
- If we change the models or their order, we will update this page and our Privacy Policy.
- The conditions of use and prohibited uses of BYVE AI are set out in our Terms of Use and Acceptable Use Policy.
Training data
- We do not collect or hold training data. We do not train or fine-tune models or create synthetic data. We do not use users' input or generated apps to train AI models.
- Information consulted during generation: when creating a new app, we may run one OpenAI web search on the app's general category and consult public product pages and app store listings as design references. Our instructions to the AI say not to use the user's full request or personal data in the search and not to copy the referenced brands. We record the search terms and the URLs consulted.
- Crawlers: we do not operate crawlers. The search above is performed through OpenAI's service.
- For each model's training data, please see its developer's published information.
Accountability
- Responsible person: the representative of By My Side Labs
- Records: we keep generation request status and token history for 365 days and recovery copies of generated results for up to 24 hours on our servers (see our Privacy Policy). We also keep the design search records (search terms and URLs consulted).
- Allocation of responsibility: our Terms of Use state that users are responsible for using, storing and sharing generated apps in compliance with laws and app store policies.
- Inquiries: we accept inquiries at the contact below and keep records of our responses.
Principle 1, disclosure: measures to protect intellectual property
| Item in the Code | Our status |
|---|---|
| Set IP protection principles, clarify responsibility, review them at least once a year and publish a summary | This page is our policy. Our representative is responsible for it, reviews it at least once a year and publishes the results here. |
| Do not infringe others' IP when using data for development and training | We do not train models. For information consulted during generation, we instruct the AI not to copy the referenced brands, as described above. |
| Use crawlers that respect access restrictions such as paywalls and robots.txt, and publish this | Not applicable: we do not operate crawlers. |
| Keep training logs for a certain period | Not applicable: we do not train models. |
| Avoid crawling pirate sites | Not applicable: we do not operate crawlers. |
| Technical measures to prevent generating infringing content | Our instructions to the AI say not to copy the brands of referenced products. For games, they say not to use the titles, names, logos or trademarked character artwork of existing works as they are, and to replace them with original names and designs. Requests that violate our Acceptable Use Policy are refused before generation. |
| Digital watermarks, C2PA and other provenance technologies | Not implemented at present: BYVE AI generates apps that run on the device (HTML and JavaScript), not image, audio or video files. |
| Tell users not to use output they believe may infringe others' IP | Our Terms of Use state that users are responsible for using, storing and sharing generated apps in compliance with laws and app store policies. |
| Provide a contact point for rights holders, clarify request requirements and keep records | We accept requests at the contact below, list the required details on this page and keep records of our responses. |
Principle 2, requests from rights holders
We respond to requests that meet the Code's requirements from anyone who is pursuing or preparing litigation, mediation, ADR or other legal proceedings to protect rights or legally protected interests in works and similar content, including a lawyer or other agent acting for them.
- What we answer: whether the URL or other information you ask about appears in the data we can check, including our design search records. Because we do not hold training data, for each model's training data we tell you the name of the model's developer.
- What to include: (1) reasons showing that you are such a requester; (2) the purpose of use and a pledge not to use the answer for any other purpose; (3) the URL or other information concerned and your specific reasons for asking.
- Fee: free
- Response time: generally within 30 days of receipt
Principle 3, requests from users
We respond to requests asking whether a URL or other information with content identical or similar to something generated with BYVE AI is included in training data.
- What we answer: the same as under Principle 2.
- What to include: (1) the generated output and the request (prompt) used to generate it; (2) the purpose of use and a pledge not to use the answer for any other purpose, including litigation, mediation or ADR; (3) the URL or other information concerned and the specific circumstances.
- Fee: free
- Response time: generally within 30 days of receipt
Contact
By My Side Labs (BYVE AI)
Email: mark@bymysidelabs.com
Please put "Principles-Code inquiry" in the subject line.