COMMON QUESTIONS.DIRECTANSWERS.

Token Boat is a unified AI API gateway. One account, API key, and compatible endpoint can access multiple models, while the console provides usage, request logs, and account pricing.

Create an API key in the console, confirm a current model in the catalog, and follow a Curl, Python, or JavaScript example in the developer docs. Start with a small test request to verify the model ID, endpoint, and response format.

Compatible endpoints generally work with the OpenAI SDK by changing the base URL, API key, and model ID. Parameters and endpoints vary by model, so verify the model profile and documentation before production use.

Start with the workload—text, reasoning, code, image, or video—then compare context, output limits, latency, and price. Rankings show anonymous aggregate usage; they are not a quality benchmark or endorsement.

Public pages compare published billing components. Discounts, group multipliers, channel adjustments, exchange rates, and the price applied to your account are shown after sign-in. Confirm account pricing before large workloads.

Text models commonly bill input and output tokens. Media models may bill image count, resolution, video duration, quality, or task count. Each model profile lists the components currently available for public display.

Keep keys only in trusted server-side environments—never browser code, mobile bundles, public repositories, screenshots, or support tickets. Separate keys by application, limit access, rotate regularly, and revoke an exposed key immediately.

Record the request time and timezone, model ID, endpoint, HTTP status, and Request ID. Check service status and account balance, then use console request logs. Never include a complete key or sensitive prompt in a support request.

A 429 usually indicates a request-rate, token, concurrency, or account-policy limit. Respect Retry-After, use exponential backoff with jitter, and reduce concurrency. Exact limits may vary by account, model, and current policy.

The gateway sends the content required to complete a call to the upstream provider for the selected model. Changing models may change the data processor, so do not submit personal data, credentials, or sensitive business content unless necessary.

The console may show request time, model, token or media usage, status, latency, Request ID, and error diagnostics. Body logging depends on service configuration and applicable rules; retention is based on data category and legal, financial, and security needs.

Check the status page for public incidents. Use the support center for account, sign-in, billing, or request-specific issues and include a Request ID. Email support@quantumnous.com for business, terms, privacy, or security enquiries.