Legal

Terms of Service

Last updated: April 2026 · Beta

These Terms govern your use of Taiva, a macOS application developed and distributed privately for beta testing. By installing or using Taiva you agree to these Terms.

1. What Taiva is

Taiva is a macOS menu bar application that uses the Anthropic Claude AI model to help you understand, summarise, and ask questions about documents, emails, and web pages open on your computer. Its features include:

Ask AI: Ask free-form questions about the active document.
Document Summary: Generate a one-click summary of a PDF, webpage, or email.
Text Selection Tools: Explain, summarise, translate, rewrite, or fix grammar for highlighted text anywhere on your Mac.
Screenshot Analysis: Capture any area of your screen and ask Claude about it.
Selection Analyzer: A floating panel (Cmd+Shift+A) for quick actions on selected text.

2. Beta status

Taiva is pre-release beta software shared privately with invited testers.

The app is free to use during the beta period.
Features may change, break, or be removed without notice.
The developer reserves the right to end the beta and discontinue access at any time.
Do not use Taiva for anything where accuracy or reliability is critical.

3. Permitted use

You may use Taiva for your own personal, non-commercial purposes during the beta. You agree not to:

Reverse-engineer, decompile, or extract the source code or embedded API keys from the application binary.
Use Taiva to process documents you do not have the right to access or share.
Use Taiva to generate content that is illegal, harmful, or violates Anthropic's usage policies.
Redistribute or resell Taiva or any part of it.
Attempt to circumvent or interfere with Taiva's local HTTP server (localhost:27134) or any other component.

4. Third-party AI service (Anthropic)

Taiva's AI features are powered by the Anthropic Claude API. When you use any AI feature, your document text, selected text, screenshots, and questions are sent to Anthropic's servers. This is necessary for the app to function.

Your use of the AI features is also subject to Anthropic's Terms of Service and Privacy Policy.
The developer does not control how Anthropic stores or processes the data it receives.
Do not use Taiva to submit documents containing sensitive personal data, trade secrets, confidential legal or medical information, or anything you would not want transmitted to a third-party AI service.

Note: Section 4 should be reviewed once Anthropic's enterprise data processing terms are considered. If users are likely to submit regulated data (e.g. HIPAA, GDPR special category data), additional contractual protections are needed before public launch.

5. Permissions Taiva requests

Accessibility: Detects text you have selected in any application so it can offer inline AI actions. Taiva does not log or transmit your selections unless you explicitly trigger an AI action.
Automation (AppleScript): Reads the content of the document currently open in Preview, Mail, Pages, Word, Notes, Safari, Chrome, and other supported apps — only when you explicitly click "Scan."
Screen Recording: Captures a portion of your screen when you use the Screenshot tab and drag to select an area — only when you explicitly initiate a capture.

You can revoke any permission at any time in System Settings → Privacy & Security without uninstalling the app.

6. AI-generated content — no professional advice

Taiva's responses are generated by an AI model and may be inaccurate, incomplete, or out of date. Nothing Taiva produces constitutes legal, medical, financial, tax, or other professional advice. Always verify AI output before acting on it.

7. Subscription and payment

The subscription system is not active during the beta. No payments will be taken. If a paid tier is introduced in the future, updated Terms will be provided before any charge is made.

Note: Subscription terms (refund policy, auto-renewal disclosures, cancellation rights) must be drafted before any paid launch, and must comply with applicable consumer protection law in your jurisdiction and the jurisdictions of your users.

8. Intellectual property

Taiva, its design, code, and branding are the property of the developer. The AI responses generated by Taiva are produced by Anthropic's Claude model; ownership of AI-generated content is subject to applicable law and Anthropic's policies.

9. Disclaimer of warranties

Taiva is provided "as is" and "as available," without any warranty of any kind. The developer expressly disclaims all warranties, including merchantability, fitness for a particular purpose, and non-infringement.

10. Limitation of liability

To the maximum extent permitted by law, the developer shall not be liable for any indirect, incidental, special, or consequential damages, including loss of data or business interruption, arising out of your use of Taiva.

Note: Liability caps and consequential damage disclaimers are not enforceable in all jurisdictions. Review for compliance with EU, UK, and Australian consumer law before public launch.

11. Changes to these terms

The developer may update these Terms at any time. Continued use of Taiva after changes are communicated constitutes acceptance of the updated Terms.

12. Contact

Questions? Reach out directly to the developer who shared Taiva with you.