# Router One Reseller & Upstream Facts

Last updated: 2026-09-05

Router One is a managed multi-model upstream you can build your own LLM API business on. The facts below describe only what ships today.

## Available today (only verified facts)

- Channel/upstream: Router One can be added as an OpenAI-compatible channel/upstream to relay software (one-api, new-api) or your own app with one API key — apps and SDKs use base URL https://api.router.one/v1; a one-api / new-api channel of type OpenAI takes the bare host https://api.router.one because the relay appends the /v1 path itself, while one-api's separate OpenAI-compatible channel type takes https://api.router.one/v1 (it strips /v1 from the path before appending /chat/completions)
- One endpoint, many models: a single OpenAI-compatible endpoint reaches 30+ models (GPT, Claude, Gemini, Grok, and others in the Router One catalog)
- Per-key controls: multiple API keys per account, each with a per-key spend limit (maxSpend), request rate limit (rateLimit), and tokens-per-minute limit (tokenLimitTpm)
- Usage data for your billing: per-request usage data (model, provider, tokens, cost, latency, status) is available for the operator's own billing and markup
- Posted-rate pricing: the pay-as-you-go token line uses posted model rates, with checkout-visible FX/channel fees kept separate; your customer margin is whatever you charge on top in your own portal
- Exact-model behavior: when a request names an exact model, retries stay on that requested model and may use another eligible provider route; model="auto" uses server-configured candidates
- Routing and retry visibility: eligible upstream failures can try another candidate under gateway policy; the customer trace records the final model/provider and request metrics, while internal failed attempts are correlated operationally by request ID
- Data handling: Router One does not retain prompt or model-response bodies from direct API calls. It records request metadata for billing, usage tracking, and troubleshooting. Playground saves conversation history so users can revisit and continue their chats.
- Referral program: pays 5% commission on referred users' top-ups (affiliate income, separate from running your own API business). Commission is credited to wallet balance and spendable on model usage; withdrawals are not supported today.

## Not available today

Router One is upstream infrastructure, not a white-label storefront. The following are NOT provided, so resellers bring their own:

- White-label branding
- Sub-accounts / child accounts
- Organization/team structure with roles
- Configurable markup multipliers
- Project-level (vs key-level) allocation
- Wholesale / volume discount tiers

Resellers bring their own portal, branding, pricing, and customer billing; Router One provides the managed multi-model upstream, per-key controls, and usage data.

## Canonical reseller pages

- LLM API reseller: https://router.one/llm-api-reseller
- White-label LLM API (boundary page): https://router.one/white-label-llm-api
- API relay station (one-api / new-api channel setup): https://router.one/llm-api-relay-station
- Wholesale LLM API (no volume discount tiers): https://router.one/wholesale-llm-api
- Referral program: https://router.one/referrals
- LLM API gateway: https://router.one/llm-api-gateway
- Company facts: https://router.one/facts/company.md