Operating agreement

This Operating Agreement explains the day-to-day rules for using the APIArc model gateway, account console, and related APIs. It supplements the Service Agreement and should be read with the Privacy Policy. If these documents conflict, the Service Agreement controls.

APIArc is powered by OpacityMade LLC.

Independent AI providers may impose their own terms, usage policies, and technical limits on the routes you use. Those provider rules continue to apply.

You are responsible for activity carried out through your account and API keys. Limit access to people you authorize, keep credentials confidential, and use separate keys where appropriate for different applications or environments.

If a key or account may be compromised, revoke the affected key or change the relevant credentials promptly and contact us. We may require verification before making account or billing changes.

Send requests using the model identifiers, endpoints, and parameters shown in the current documentation or console. Supported capabilities, context limits, and provider restrictions vary by model. You are responsible for checking that a route fits your use case before relying on it.

APIArc may select an available upstream route for the model you request based on compatibility, capacity, and service health. Providers, routes, model versions, and limits may change. A request can fail or be delayed when a provider is unavailable, a limit is reached, or a safety control applies.

The model catalog and console display the applicable billing units and prices. Usage is measured using APIArc and provider telemetry; token or unit counts may differ slightly between systems. Charges and account balances are recorded in the console.

Requests may be paused or rejected when a balance, spending limit, or payment condition prevents further use. Purchases, auto-recharge, refunds, taxes, and promotional credits are governed by the Service Agreement, the applicable checkout terms, and any specific promotion rules. Report a suspected billing error within 30 days so we can investigate it.

We may perform maintenance, update integrations, apply rate limits, or temporarily disable a route to protect the service or respond to provider changes. Unless a separate written service-level agreement applies, we do not guarantee uninterrupted availability or a particular model route.

For an operational issue, send the request ID, approximate time, model, endpoint, and error details to support. Do not send API keys, passwords, payment details, prompts, or model outputs by email.

You must comply with the Service Agreement, applicable law, and the rules of the provider serving your request. Do not use the service to bypass safeguards, interfere with systems, or conduct prohibited or abusive activity.

We may throttle, block, or suspend activity when reasonably necessary to address security threats, abuse, unpaid charges, legal requirements, or provider restrictions. We will give notice when practical, consistent with the Service Agreement.

APIArc processes request content to route it to the selected AI provider and return the response. If prompt auditing is enabled for a request, submitted prompt text may also be sent to a configured audit endpoint and stored in a security audit record. We do not use model content to train models.

We retain operational metadata such as request time, model, token or unit counts, latency, status, and calculated cost. The Privacy Policy explains audit records, retention, and your choices. The selected provider processes request content under its own policies.

We may revise this Operating Agreement as the service, providers, or legal requirements change. We will post the revised text and update the effective date; material changes will receive additional notice where the Service Agreement or law requires it.

For questions about how APIArc operates, contact us at the address below.

contact@apiarc.ai