Terms of service.

The framework for using MegaRider.

Effective 7 September 2026. Contact: hello@megarider.pro.

Access and agreement

MegaRider offers an inference API to developers and integration partners. The operator’s legal identity, contracting address, jurisdiction, prices and payment terms are specified in the service agreement before paid access. These website terms describe the general service; a signed agreement prevails where it differs. No payment or subscription is created by visiting this site or submitting an access application.

Infrastructure and external dependencies

The service uses a MegaRider-operated virtual server, a Caddy TLS reverse proxy, a Node.js API gateway and SQLite usage/account records. Model computation is supplied through OpenBroker and the Gonka network. We manage the API delivery layer; we do not claim exclusive ownership or control of the underlying GPU fleet. Our mail system uses Postfix, Dovecot and Rspamd through Docker Mailserver, with Roundcube for webmail. Third-party infrastructure availability and data handling are outside our complete control. See the privacy notice for details.

Your responsibilities

You must have authority to use the service and provide the submitted content. Keep access credentials confidential, respect assigned limits, and comply with our acceptable use policy, applicable law and any terms attached to the selected model. You are responsible for your users and any tools your application executes.

The service

We operate the API gateway and use external inference infrastructure. Model availability and features can change. Limits, supported parameters and errors are described in the documentation. We may reject requests or suspend access to protect the service, address non-payment, investigate abuse or meet legal obligations.

Output and intellectual property

You retain rights you hold in submitted content. You authorize the processing needed to supply the service. Outputs may be inaccurate, non-unique or subject to third-party rights. Review them before use, especially for decisions affecting people. No rights beyond those permitted by the applicable model license are granted.

Charges and reconciliation

Paid use begins only under agreed rates and billing terms. Token counts come from inference usage reports. Input and output can be priced separately; reasoning tokens are included in reported output where supplied by the model. Retries can create additional usage. Missing, failed or interrupted usage records require reconciliation and are not automatically treated as successful billable requests by our gateway. Report disputed usage with request IDs through support.

Availability and responsibility

Unless a separate written agreement states otherwise, MegaRider is provided AS IS and AS AVAILABLE. We do not promise a particular uptime, response time, throughput, model behavior, output accuracy, commercial result or fitness for your specific purpose. The service may be delayed, interrupted or unavailable because of capacity, maintenance, network conditions or inference-provider failures. An evaluation, example or latency measurement does not create a service-level commitment. Your expectations or intended use do not by themselves establish a guarantee; evaluate the service with your own workload before relying on it. You remain responsible for reviewing output, securing integrations, setting spending limits and maintaining appropriate fallback and human review. Any agreed remedies and allocation of liability are governed by the applicable service agreement and law. Nothing here removes mandatory consumer rights or excludes liability that cannot lawfully be excluded, including fraud or personal injury caused by negligence where applicable.

Ending access and changes

You may stop using the service and request key revocation through support. Accrued obligations under a paid agreement survive termination. Material service or contractual changes will be communicated to active partners through their agreed contact channel.