Terms of Service

These Terms set forth the conditions for using Shirabe AI, a generative AI research workflow service operated by MU Inc.

Users must agree to these Terms before using the service.

Some features are available without registration as Guest usage, while account-based features are provided under Free, Basic, and Pro plans.

Guest usage may be limited by model, skill, file upload, connected services, and monthly usage count.

The Company may refuse or suspend registration when false information is provided, these Terms are violated, or the Company otherwise determines that use is inappropriate.

Paid plan fees are shown on the pricing page. Payment is made by credit card through the payment method designated by the Company.

Subscription access is provided after payment is completed. Unless otherwise stated, paid plans are provided on a monthly or yearly subscription basis.

Users must not violate laws or public order, infringe third-party rights, submit unlawful or harmful content, interfere with service operation, access systems without authorization, collect information about other users, impersonate others, or provide benefits to antisocial forces.

Users must not use generated outputs in a way that misleads others regarding medical, legal, financial, academic, or professional facts.

The Company may suspend all or part of the service without prior notice for maintenance, system failure, external service failure, force majeure, security reasons, or other operational necessity.

The Company may restrict use or delete accounts when a user violates these Terms or the Company determines that continued use is inappropriate.

Generated text, slides, spreadsheets, videos, code, and other outputs may contain errors, omissions, outdated information, or content requiring rights clearance.

Users are responsible for reviewing outputs, checking source materials, verifying citations, and determining whether the output may be used or published.

For medical or clinical matters, users must not rely solely on AI outputs and must consult a physician or other qualified professional.

The Company may store and use account information, chat history, operation logs, uploaded-file metadata, generated-output metadata, and connected-service authorization data only as necessary to provide, maintain, secure, and improve the service.

Input and output data handled by Claude Managed Agents are processed and stored in Anthropic's managed cloud environment, which is separated from the application database.

The Company does not use input or output data to train, retrain, or fine-tune foundation models.

The Company does not guarantee the accuracy, completeness, usefulness, fitness for a particular purpose, legality, non-infringement, or availability of the service or generated outputs.

The service operates AI agents on Claude Managed Agents. Based on user instructions or autonomous decisions, AI agents may edit or delete files, send requests to connected external services, and perform other destructive actions.

The Company is not liable for any file deletion, data loss, impact on connected external services, or other damages arising from AI agent actions. Users are responsible for taking their own backups, managing permissions, and reviewing operations before authorizing them.

The Company may change these Terms or the service contents when necessary.

These Terms are governed by Japanese law. Disputes arising from the service shall be subject to the exclusive jurisdiction of the court having jurisdiction over the Company's head office location.