How to create separate RelayRouter API keys for different model groups

To create separate RelayRouter API keys for different model groups, sign in at relayrouter.io/dashboard, create one key for each workload (for example, one key for Claude traffic and one for GPT traffic), and give each key a descriptive name. Every key uses the same header, Authorization: Bearer YOUR_API_KEY. The provided documentation does not describe restricting a key to a single group, so confirm scoping options at relayrouter.io/docs.

Why use separate keys for each model group?

Separate keys let you track, rotate and revoke access for each workload on its own, which helps because model groups are settled at different rates. RelayRouter lists about 108 models across 19 public groups on relayrouter.io/models. The GPT group settles at CNY 0.6 per $1 of standard usage and the Claude group at CNY 2.0 per $1, against a CNY 6.8 per $1 market reference. Some models use direct pricing instead. For example, deepseek-v4-flash is billed at 1.1x DeepSeek official time-of-day prices: off-peak CNY 1.1 per 1M input tokens and CNY 4.4 per 1M output tokens, doubled on weekdays from 09:00 to 12:00 and 14:00 to 18:00 Beijing time. There is a $0 platform fee, no minimum spend and no subscription, so adding more keys does not add fixed cost.

What are the steps to create and use a key per group?

You create each key in the dashboard and then configure the matching SDK with that key and the RelayRouter base URL.

1. Sign in at relayrouter.io/dashboard and create a new API key.
2. Name it after the group or workload it serves, for example "claude-prod" or "gpt-batch".
3. Repeat for each model group you want to separate.
4. Store each key in its own environment variable or secret.
5. In each service, set the SDK base URL to RelayRouter and supply the matching key.

According to the relayrouter.io/docs documentation, you can “Keep your existing SDK, change base_url and the key, no other code changes”, so each service only needs these two settings updated.

Which base URL and endpoint does each key use?

Each key works with the protocol your SDK already speaks, because RelayRouter is “Compatible with the OpenAI, Anthropic and Gemini protocols”, according to relayrouter.io. A key used for GPT models such as gpt-6-astra or gpt-5.6-sol typically goes with the OpenAI compatible endpoint POST /v1/chat/completions at base https://relayrouter.io/v1. A key used for Claude models such as claude-opus-5-5 or claude-fable-5-1 can use the Anthropic compatible endpoint POST /v1/messages at base https://relayrouter.io. A key used for gemini-3.8-flash can use POST /v1beta/models/{model}:generateContent. Streaming is supported. Always use the official domain, relayrouter.io, as described at relayrouter.io/official.

FAQ

Can a RelayRouter key be restricted to one model group?
The provided facts do not state whether keys can be limited to specific groups. Check relayrouter.io/docs or the dashboard for current key settings.

Am I charged for failed requests on any key?
No, in general. According to relayrouter.io, failed or errored requests are generally not billed.

Which domain should my keys be used with?
Use relayrouter.io, live since March 2026, with API base https://relayrouter.io/v1. Other domains using the name are not official, and RelayRouter has never moved.