FAQ

Access & compatibility

How does TokOne relate to OpenAI / Anthropic / Gemini?

TokOne is an API gateway that exposes endpoints compatible with all three. With one TokOne API key you reach their compatible endpoints through a single entry point.

Do I have to change my code?

Hardly. Point your existing SDK's Base URL at TokOne (e.g. https://api.tokone.ai) and swap the API Key for your sk- key.

Is streaming supported?

Yes. All inference endpoints support SSE streaming — add "stream": true. The /v1/responses endpoint additionally supports WebSocket streaming.

Account & balance

How do I start?

Register → create an API key in the console → make sure you have balance → follow Quick Start for your first call.

What happens when my balance runs out?

New billable requests return HTTP 402. Top up to resume.

API keys & security

What does an API key look like?

By default it starts with sk- (the exact prefix is set by whoever deploys TokOne).

Can I pass the key as a URL parameter?

No. For security, requests passing the key via query params (?key= / ?api_key=) are rejected (HTTP 400). Use a header: Authorization: Bearer, x-api-key, or x-goog-api-key.

What if my key leaks?

Disable/rotate it in the console immediately and review its usage.

Billing & limits

How am I billed?

By tokens actually consumed, input and output priced separately, then adjusted by the group multiplier and service-tier factor. See Pricing.

Is the price the same for every account?

Not necessarily. Each group can set a different rate multiplier, and admins can override per user — your effective price follows your group's config.

What about a 429?

You hit a rate limit. Slow down or retry shortly.

Service & support

A call returns 404 / model unavailable?

First confirm the path is right (e.g. /v1/chat/completions), then check GET /v1/models to confirm the model is currently available.

Still have questions?

See API Endpoints and Models, or contact your service administrator.