Terms & Conditions

Effective date: 14 May 2026  ·  Last updated: 14 May 2026

Please read these Terms and Conditions ("Terms") carefully before using the AgentHeaven service operated by Black Salt Kft. ("we", "us", or "our") at agentheaven.ai, registered at 1077 Budapest, Baross tér 14, Hungary (company registration no. 01 09 450602, VAT no. HU13853619).

By subscribing to our service or accessing our website, you agree to be bound by these Terms. These Terms incorporate our Acceptable Use Policy, Data Processing Addendum where applicable, Sub-processors page, Security Measures, Privacy Policy, and Cookie Policy. If you do not agree with any part of these Terms, you may not use our service.

1. The Service

AgentHeaven provides a managed artificial-intelligence digital-employee service ("the Service") that, on your behalf:

• Configures a digital employee to perform a defined role (e.g., research, inbox triage, customer support, repetitive task automation, internal knowledge assistance, ops/back-office tasks);
• Hosts the agent on VPS infrastructure controlled by us, using an agent harness and a model-agnostic routing layer over optional self-hosted open-source models and selected third-party LLM providers;
• Absorbs all underlying compute and LLM costs as part of the flat monthly subscription;
• Maintains a customer-specific long-term memory/wiki for the agent, containing relevant business information, workflows, preferences, and approved learning from sessions;
• Continuously tunes the agent based on your feedback, session logs, the customer-specific memory/wiki, and our team's ongoing review, within the tuning-hours allowance of your subscription tier.

The Service is configured by us following an onboarding consultation with you. You provide us with access to the tools the agent needs (e.g., CRM, email, helpdesk, document store) and a description of the work to be done. We process source documents and connected-tool content during agent sessions, but do not store your source documents by default. We do store session and activity logs, and we store the customer-specific memory/wiki needed for your agent to improve over time. We manage the digital employee on an ongoing basis.

2. Eligibility

You must be at least 18 years old and have the legal authority to enter into these Terms on behalf of yourself or your business. By using the Service, you represent and warrant that you meet these requirements.

3. Your Account and Responsibilities

By subscribing to the Service, you agree to:

• Provide accurate, complete, and up-to-date information during onboarding and throughout your subscription;
• Grant us the necessary access and permissions for the tools the digital employee is configured to operate;
• Keep us informed of any changes to your business processes, runbook, or preferences that may affect the agent's output;
• Review and correct the customer-specific memory/wiki if you become aware that it contains inaccurate, outdated, confidential, or inappropriate information;
• Review agent activity periodically and notify us promptly if you identify any output that is inappropriate, inaccurate, or off-brief;
• Comply with the terms of service of any third-party platforms the agent connects to on your authorisation.

4. Content and Output Responsibility

While we configure and operate the digital employee on your behalf, you retain full responsibility for the output of the agent, including any content, messages, decisions, or actions performed by the agent on your authorisation.

You agree that you will not instruct us to use the digital employee to create or transmit content that:

• Is unlawful, defamatory, harassing, discriminatory, or otherwise harmful;
• Infringes on the intellectual property rights of any third party;
• Is misleading, deceptive, or in violation of advertising standards;
• Violates the community guidelines or terms of service of any platform the agent is connected to;
• Promotes illegal products, services, or activities.

We reserve the right to refuse to operate the agent in a manner that we reasonably believe violates the above standards, applicable law, our Acceptable Use Policy, or the acceptable-use policies of the underlying LLM providers.

5. Acceptable Use

In addition to Section 4, you may not use the Service:

• To send mass unsolicited communications (spam) of any kind;
• To attempt to circumvent rate limits, security controls, or fair-use limits;
• To facilitate phishing, fraud, identity theft, or other malicious activity;
• To generate content prohibited by our Acceptable Use Policy, the acceptable-use policies of underlying LLM providers (Anthropic, OpenAI, and open-source model licences), or connected-platform rules;
• To reverse-engineer, decompile, or extract the prompts, weights, runbook, or memory of any digital employee, your own or anyone else's.

Violation of this Section may result in immediate suspension or termination of the Service without refund.

6. Subscription, Billing, and Payment

The Service is offered on a monthly subscription basis at the tier you select at the time of sign-up (currently Starter (€299/mo), Specialist (€699/mo), or Workforce (€1,499/mo)). Each tier includes a defined number of human-tuning hours per month and a defined number of connected tool sources, as described on the pricing page.

• Billing cycle: Subscriptions are billed monthly in advance from the date you subscribe.
• Payment: Payment is processed via our third-party payment provider. By subscribing, you authorise us to charge your payment method on a recurring monthly basis.
• Price changes: We may change subscription prices with at least 30 days' written notice. Existing subscriptions keep their current price until renewal.
• Failed payments: If a payment fails, we will attempt to notify you. Continued failure to pay may result in suspension or termination of the Service.
• Fair use: AgentHeaven absorbs the underlying compute and LLM costs subject to fair use. We may apply reasonable rate limits per tier to prevent abuse. We will discuss limits openly with any customer before applying them.

VAT treatment: All prices are stated exclusive of Value Added Tax (VAT) unless otherwise indicated.

• Hungarian customers: Hungarian VAT at the current statutory rate (27%) will be added to all invoices.
• EU business customers (B2B): If you are a VAT-registered business in another EU member state and provide a valid VAT number, the reverse charge mechanism applies under Art. 196 of the EU VAT Directive and no VAT will be charged by us.
• EU consumer customers (B2C): VAT will be charged at the rate applicable in your EU member state of residence, in accordance with EU digital services VAT rules.
• Non-EU customers: No VAT is charged by AgentHeaven. You are responsible for any local taxes or duties that may apply in your jurisdiction.
• Invoicing: Invoices are issued in accordance with Hungarian invoicing regulations.

7. Cancellation and Refunds

You may cancel your subscription at any time by emailing us at info@agentheaven.ai. The cancellation takes effect at the end of your current billing month. There is no notice period and no exit fee.

Refund policy: Pro-rated refunds are available within the first 14 days of a new subscription if you decide AgentHeaven is not a fit. After 14 days, cancellation stops future billing but does not refund the current month.

After cancellation:

• Your digital employee is deactivated and ceases to operate;
• Tool access tokens you authorised are revoked and deleted from our systems;
• Your personal data is handled in accordance with our Privacy Policy.

We reserve the right to cancel or suspend your subscription if you breach these Terms. For non-material breaches, we will provide 14 days' written notice and an opportunity to remedy the breach. For serious or material breaches (including violations of Sections 4 or 5, or activity that creates immediate risk to AgentHeaven or third parties), we may suspend or terminate with immediate effect and without refund.

8. Service Levels, Availability, and Support

We aim to provide a reliable, continuously operating service. We target 99.0% monthly uptime for the agent orchestration layer under our direct control. We do not guarantee uptime of third-party LLM providers or external tools the agent is connected to.

The Service may be temporarily unavailable due to:

• Scheduled maintenance (we will endeavour to provide advance notice);
• Technical failures or circumstances beyond our reasonable control;
• Outages or rate limits imposed by third-party LLM providers or external tools.

Support: Support is provided via email at info@agentheaven.ai. We aim to respond within 2 business days for Starter customers and within 1 business day for Specialist customers. Workforce customers receive a response on tuning and support requests within 1 business day under our service-level commitment.

9. Intellectual Property

Your content: You retain ownership of all business information, runbook content, brand materials, operational data, and customer-specific memory/wiki content you provide to us or authorise the agent to derive from your sessions. You grant us a limited licence to use this material solely to configure, operate, secure, support, and tune your digital employee.

No cross-customer use: We do not use your business information, source documents, session content, logs, or customer-specific memory/wiki to train shared models, improve another customer's agent, create generic datasets, or develop unrelated AgentHeaven products. We do not disclose your customer-specific memory/wiki to third parties except where strictly necessary to operate your selected LLM routing mode or connected-tool workflow.

Our service: All technology, software, processes, methods, prompts, and tuning materials used to deliver the Service remain the intellectual property of AgentHeaven and its partners (including the Hermes Agent harness, licensed from Nous Research under its applicable terms). These Terms do not grant you any licence to our intellectual property beyond the right to use the Service as described herein.

AI-generated output: AgentHeaven grants you a perpetual, royalty-free, worldwide licence to use, copy, distribute, and repurpose the output produced by your digital employee in the course of your subscription. Note that AI-generated content may not qualify for copyright protection under applicable law; we make no warranty as to its copyright status.

10. Confidentiality

We treat your business information, runbook, operational data, session logs, customer-specific memory/wiki, and tool access credentials as confidential. We will not disclose this information to third parties except as necessary to operate the Service (see our Sub-processors page), as required by your selected LLM routing mode or connected-tool workflow, or as required by law.

Equally, you agree to keep confidential any proprietary information about our processes, tuning methodology, or technical methods that we share with you during the course of the Service.

11. Limitation of Liability

Our total aggregate liability to you for any claim arising under or in connection with these Terms shall not exceed the total subscription fees paid by you in the three (3) months preceding the claim.

Nothing in these Terms limits our liability for death or personal injury caused by negligence, fraud, or any other liability that cannot be excluded by law.

12. Indemnification

You agree to indemnify and hold harmless AgentHeaven, its team members, and partners from and against any claims, losses, damages, or expenses (including reasonable legal fees) arising from:

• Your breach of these Terms (in particular Sections 4 and 5);
• Output the digital employee produces on your authorisation that violates applicable law or third-party rights;
• Your violation of any third-party platform's terms of service.

13. Modifications to the Service and Terms

We reserve the right to modify or discontinue the Service (or any part of it) at any time, with reasonable notice where possible. We will notify active subscribers of any significant changes to the Service by email.

We may update these Terms from time to time. The updated version will be posted on this page with a revised effective date. Continued use of the Service after changes take effect constitutes your acceptance of the revised Terms. If you do not agree with the changes, you may cancel your subscription before the new Terms take effect.

14. Consumer Right of Withdrawal

If you are a consumer (not a business) located in the EU or EEA, you have the right to withdraw from this contract within 14 days of subscribing, without giving any reason (the "cooling-off period"), under the EU Consumer Rights Directive.

However, by requesting that we begin setting up your digital employee immediately upon subscription, you expressly consent to the Service commencing before the end of the 14-day withdrawal period. If you then exercise your right of withdrawal, you acknowledge that you will be liable to pay for the portion of the Service already delivered up to the point of withdrawal, calculated pro-rata.

To exercise your right of withdrawal, contact us at info@agentheaven.ai with the subject line "Withdrawal" within the 14-day period.

This right does not apply to business subscribers.

15. Governing Law and Disputes

These Terms are governed by and construed in accordance with the laws of Hungary. Any disputes arising out of or in connection with these Terms shall be subject to the exclusive jurisdiction of the courts of Budapest, Hungary, except where mandatory applicable consumer law in your country of residence provides otherwise.

We will always attempt to resolve any dispute amicably in the first instance — please contact us at info@agentheaven.ai before pursuing any formal action.

If you are a consumer located in the EU, you may also seek assistance through the consumer redress bodies listed by the European Commission: https://consumer-redress.ec.europa.eu/dispute-resolution-bodies. The former European Online Dispute Resolution platform was discontinued on 20 July 2025.

16. Entire Agreement

These Terms, together with our Privacy Policy, Data Processing Addendum where applicable, Acceptable Use Policy, Sub-processors page, Security Measures, and Cookie Policy, constitute the entire agreement between you and AgentHeaven regarding the Service and supersede all prior agreements and understandings, whether written or oral.

If any provision of these Terms is found to be unenforceable, the remaining provisions will continue in full force and effect.

17. Contact

For any questions about these Terms, please contact us:

---

These Terms were last reviewed on 14 May 2026. AgentHeaven reserves the right to update this document at any time. The current version is always available at agentheaven.ai/terms.