Privacy Policy

1. Scope and contact

This Policy covers the AtmoRouter website, API, dashboard and related account services. It describes AtmoRouter’s handling of information; external websites, payment services and model infrastructure may have their own data practices.

For privacy questions or requests, contact [email protected]. Do not include passwords, full API keys or unnecessary sensitive information in your message.

2. Information we process

• Account and access information: email address, account status, password verification data, sessions, API-key verification data, key labels and access settings.
• Inference content: prompts, conversation history, tool definitions, files or other content you submit, and the responses needed to fulfil your request.
• Usage and technical records: model identifier, token and cache counts, timestamps, request identifiers, response status, timing, recorded charges, price limits, source IP address and client user-agent.
• Payment and account records: credited balance, payment references, supported asset and network details, transaction identifiers, adjustments and relevant referral or payout records.
• Website and support information: page visits, device category, referrers, campaign attribution, communications and attachments you choose to send us.

Payment providers may collect information directly to process a payment. Blockchain transactions may be publicly visible and cannot be made private or deleted by AtmoRouter.

3. How information is used

We use information to create and secure accounts, authenticate requests, route inference, calculate and reconcile charges, credit payments, provide usage records, respond to support, detect abuse and comply with applicable obligations. We also use website and attribution information to understand visits and assess marketing effectiveness.

Where data-protection law applies, processing may be necessary to provide the service you request, protect legitimate operational and security interests, meet legal obligations, or act on consent where required. The applicable basis depends on the information, purpose and law involved.

4. Prompts, responses and model processing

Inference content is processed to route your request and return a response. AtmoRouter’s normal usage and billing records contain metadata rather than a stored copy of your prompt or answer. We do not use your API request content to train models.

The infrastructure fulfilling an inference request receives the content needed to perform it. Its processing location, retention and model-specific practices can differ. We do not represent that every model has identical privacy terms or guaranteed zero retention.

If you voluntarily include content in a support request, it becomes part of that support case. Avoid unnecessary sensitive data, and contact us before using the Service for a workload with special contractual, confidentiality or regulatory requirements.

5. Service providers and disclosures

Information may be processed by services that support model inference, hosting, network delivery, payments, transactional communications, security, analytics and advertising measurement. We disclose information appropriate to the function being performed. We may also disclose information where required by law, to investigate abuse or fraud, or to protect the Service and its users.

These services may operate in countries other than your own. Contact [email protected] for questions about a particular processing arrangement or request. Applicable requirements for international processing and disclosures still apply.

6. Cookies, browser storage and marketing

Essential session cookies keep you signed in. Browser storage also supports account display, theme preferences and referral or campaign attribution. A temporary tab identifier can help distinguish visits. Blocking essential cookies or clearing storage can affect sign-in and saved preferences.

AtmoRouter records first-party website analytics, including page paths, referrers, device categories and campaign information. Its page-view analytics store a salted hash of the IP address rather than the raw address. This does not mean that account, security or API usage records never contain an IP address.

Advertising measurement integrations, when enabled, may receive events such as registration or a confirmed payment, event identifiers, value and currency, browser or device information, IP address and a normalized, hashed email address for event matching. A hashed email remains potentially linkable information; it is not a promise of anonymity. API prompts, responses, passwords and API keys are not included in these advertising events.

Browser tracking protections, cookie controls and content blockers may limit some browser-side tracking. They do not necessarily stop server-side event processing. Contact [email protected] to ask about or object to marketing-related processing where applicable. We will assess requests and any consent requirements under the law that applies.

7. Security and retention

Our safeguards include password hashing, hash-based verification of API keys and sessions, access controls, transport protection and operational security checks. Full API keys are shown at creation and cannot be recovered from the normal key-management interface. No online service can guarantee absolute security.

We retain account and usage information to provide the account, maintain billing integrity, investigate abuse and resolve disputes. Financial, security and legal records may need to remain after account closure. Routine raw website page-view records are scheduled for aggregation after approximately 90 days; daily aggregate counts can remain longer.

Retention depends on the record and purpose. Deletion requests do not erase public blockchain records, records independently held by an external service, or information we must retain under applicable obligations.

8. Your access, export and deletion options

You can manage or revoke API keys and use account settings to export available account data or request deletion. Follow the confirmation steps and any pending-deletion period shown in the interface. A refund request is handled separately under the Refund Policy.

Depending on applicable law, you may have rights to access, correct, delete or receive information, restrict or object to processing, withdraw consent, or complain to a competent authority. Send requests to [email protected]. We may verify your identity and retain information where a lawful exception applies. Withdrawing consent does not invalidate processing already lawfully performed.

9. Children and sensitive information

AtmoRouter is intended for people able to use a developer service lawfully and enter the required agreement. It is not directed at children. If you believe a child’s personal information has been provided without the required permission, contact us so we can review it.

Do not submit sensitive personal information or confidential material unless you have the necessary rights and have assessed the handling of the selected model and your own application.

10. Updates and contact

We may update this Policy as the Service and its practices change. The updated date identifies the current version; material changes will be communicated through the Service or account communications where appropriate. We will not treat an updated notice as removing rights protected by applicable law.

Questions about your information can be sent to [email protected]. Service terms are described in the Terms of Service.