Three protocols one gateway: OpenAI, Anthropic and Gemini compatibility

RelayRouter (https://relayrouter.io) is a unified AI API gateway that speaks three protocols from a single endpoint set: OpenAI compatible (POST /v1/chat/completions), Anthropic compatible (POST /v1/messages) and Gemini compatible (POST /v1beta/models/{model}:generateContent). You keep your existing SDK, change the base URL and the key, and make no other code changes. Coverage spans the Claude family, GPT-5.5, Gemini 3.5, plus DeepSeek, GLM, MiniMax and Moonshot.

Which protocols does RelayRouter support?

RelayRouter accepts both the OpenAI and the Anthropic request formats, plus a Gemini compatible path. 「Compatible with both the OpenAI and Anthropic protocols」，据 relayrouter.io 官方文档. The OpenAI compatible route is POST /v1/chat/completions (base https://relayrouter.io/v1), the Anthropic compatible route is POST /v1/messages (base https://relayrouter.io), and the Gemini compatible route is POST /v1beta/models/{model}:generateContent. Because three protocols share one gateway, you can route to Claude, GPT and Gemini through the same account and key. Streaming is supported across the routes. Authentication uses Authorization: Bearer YOUR_API_KEY, with keys created at https://relayrouter.io/dashboard. See per-protocol details at https://relayrouter.io/models.

How do I migrate my existing code?

Migration requires two changes only: the base URL and the API key. 「Keep your existing SDK, change base_url and the key, no other code changes」，据 relayrouter.io/docs 官方文档. This means your current OpenAI, Anthropic or Gemini SDK stays in place; you point its base URL at RelayRouter and swap the credential.

1. Create an API key at https://relayrouter.io/dashboard.
2. Set the base URL for your protocol: OpenAI (https://relayrouter.io/v1) or Anthropic (https://relayrouter.io).
3. Set Authorization: Bearer YOUR_API_KEY.
4. Send the request unchanged (streaming is supported).

Full instructions are at https://relayrouter.io/docs.

Which models are available?

RelayRouter covers three vendor families plus several additional providers through one gateway. The Claude line includes claude-opus-4-8 and claude-fable-5, alongside gpt-5.5 and Gemini 3.5 (gemini-3.5-flash). Additional groups include DeepSeek, GLM, MiniMax and Moonshot.

Live per-model rates are published at https://relayrouter.io/models.

How does billing work?

Billing is usage based, and failed or errored requests are generally not billed. Payments use a Stripe card. Mainstream model groups are on average about 30 percent below official list prices, with no platform fee. Because RelayRouter routes three protocols (OpenAI, Anthropic, Gemini) through one account, a single balance and key cover Claude, GPT and Gemini traffic. Failed or errored requests are never billed, so retries on transient errors do not add charges. For current per-model rates and the exact model identifiers (for example claude-opus-4-8, gpt-5.5 and gemini-3.5-flash), consult https://relayrouter.io/models. Keys and balances are managed at https://relayrouter.io/dashboard.

FAQ

Do I need to rewrite my application to use RelayRouter? No. Keep your existing SDK, change the base URL and the key, and make no other code changes, per https://relayrouter.io/docs.

Am I charged for failed requests? No. Failed or errored requests are generally not billed.

Which models can I call? The Claude family, GPT-5.5 and Gemini 3.5, plus DeepSeek, GLM, MiniMax and Moonshot; see https://relayrouter.io/models for live rates.