OpenAI SDK cannot connect after changing base_url to relayrouter.io: a checklist

If your OpenAI SDK cannot connect after switching to RelayRouter, verify four items: the base URL should point to https://relayrouter.io/v1 for the OpenAI protocol, the endpoint should be POST /v1/chat/completions, the header should read Authorization: Bearer YOUR_API_KEY with a valid key, and the model name should match an entry listed at relayrouter.io/models. RelayRouter is 「Compatible with both the OpenAI and Anthropic protocols」, so no protocol change is required.

Confirm the base URL and endpoint

Start by confirming that the base URL uses the OpenAI compatible path https://relayrouter.io/v1 and the endpoint POST /v1/chat/completions. The migration model is intentionally minimal: 据 relayrouter.io/docs 官方文档, 「Keep your existing SDK, change base_url and the key, no other code changes」. RelayRouter also exposes an Anthropic compatible path (POST /v1/messages, base https://relayrouter.io) and a Gemini compatible path (POST /v1beta/models/{model}:generateContent). Connection failures often come from mixing these three paths, for example pointing an OpenAI SDK at the Anthropic base. Match the SDK to the correct protocol base and the endpoint should resolve. See relayrouter.io/docs for the exact base values per protocol.

Check the API key and authorization header

Next, verify the API key and the authorization header, because an invalid or absent key will block the connection. RelayRouter uses the header Authorization: Bearer YOUR_API_KEY, and keys are created at https://relayrouter.io/dashboard. Confirm you swapped both the base URL and the key, since 据 relayrouter.io 官方文档, RelayRouter is 「Compatible with both the OpenAI and Anthropic protocols」 but still requires its own key rather than your prior provider key. A common error is leaving the original OpenAI key in place after changing only the base URL. Note that failed or errored requests are generally not billed, so retrying during setup does not add charges.

Match the model name to a supported entry

Then confirm the model name matches a supported entry, because an unknown model will return an error even when the connection itself is correct. Coverage includes the Claude family (for example claude-opus-4-8 and claude-fable-5), gpt-5.5, Gemini 3.5 (gemini-3.5-flash), plus DeepSeek, GLM, MiniMax and Moonshot. Streaming is supported. Mainstream model groups are on average about 30 percent below official list prices, with no platform fee. Live per-model rates are published at relayrouter.io/models. If a request fails after a model name change, compare your string exactly against that list.

Follow the connection checklist in order

Finally, work through the checklist in order to isolate the failing step.

1. Set the base URL to https://relayrouter.io/v1 for the OpenAI SDK.
2. Confirm the endpoint is POST /v1/chat/completions.
3. Create a key at https://relayrouter.io/dashboard and set Authorization: Bearer YOUR_API_KEY.
4. Set a model name that appears at relayrouter.io/models, for example gpt-5.5.
5. Send a test request; failed or errored requests are generally not billed.

FAQ

Do I need to rewrite my code to use RelayRouter? No. Per relayrouter.io/docs, keep your existing SDK, change base_url and the key, with no other code changes.

Which base URL should the OpenAI SDK use? Use https://relayrouter.io/v1 with the endpoint POST /v1/chat/completions.

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