Company

One gateway, every model.

RouterPlex exists because using more than one AI provider is unreasonably painful. Every provider has its own SDK, its own key, its own billing page, its own quota system — and the best model for your task changes every few months. Teams end up locked into whichever provider they integrated first, not whichever model is actually best.

RouterPlex collapses that mess into one account: OpenAI- and Anthropic-compatible endpoints, a single prepaid balance, and 40 chat and image models from 14 providers behind it. Change one string in your code and you've switched models. That's the whole product.

01What we believe

  • No lock-in. The API is OpenAI-compatible on purpose — leaving RouterPlex should be as easy as joining.
  • Honest pricing. Models are billed at the vendor's public list price, per token, from a prepaid balance. No seats, no minimums, no markup games — and subscriptions are optional bonus-credit bundles, never a requirement.
  • Hard limits, not surprises. Every key can carry a budget, a model allowlist, and rate limits. Your balance can never go negative.
  • Privacy by scope. We store usage metadata, not your prompt content, and we never train on your data.

02The state of things

RouterPlex is young and independently built. We launched in July 2026 and now serve 40 models across 14 providers, per-token billing, scoped keys, a playground, and card & crypto top-ups. The roadmap is driven by what users ask for — if something is missing for your use case, tell us and there's a good chance it ships quickly.

03Get in touch

Questions, feedback, model requests, partnership ideas: support at routerplex.com

Or just try it — create an account and top up from $5 to send your first request in under a minute.