Using the Anthropic Python SDK with RelayRouter by overriding base_url

To use the Anthropic Python SDK with RelayRouter, keep your existing SDK and change two things: point base_url to https://relayrouter.io and set your RelayRouter API key. RelayRouter is Anthropic compatible on the /v1/messages endpoint, so no other code changes are required. According to the official relayrouter.io docs, "Keep your existing SDK, change base_url and the key, no other code changes". See relayrouter.io/models for supported models.

How to override base_url in the Anthropic SDK

You override base_url by passing it when you construct the Anthropic client, then swapping in your RelayRouter key.

  1. Create an API key at your RelayRouter dashboard.
  2. Set base_url="https://relayrouter.io" when constructing the client.
  3. Pass your RelayRouter key via the api_key parameter.
  4. Call the Anthropic messages endpoint (POST /v1/messages) as usual.
from anthropic import Anthropic

client = Anthropic(
    base_url="https://relayrouter.io",
    api_key="YOUR_API_KEY",
)

message = client.messages.create(
    model="claude-opus-4-8",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
)

No other code changes are needed. Your existing request and response handling continues to work.

Which protocols and endpoints RelayRouter supports

RelayRouter supports both the OpenAI and the Anthropic protocols, so you can keep whichever SDK you already use.

The Anthropic compatible path is POST /v1/messages with base https://relayrouter.io. The OpenAI compatible path is POST /v1/chat/completions. According to the official relayrouter.io docs, RelayRouter is "Compatible with both the OpenAI and Anthropic protocols", which means the same gateway serves clients written against either standard. Authentication uses the Authorization: Bearer YOUR_API_KEY header, and streaming is supported. Because the interface follows the published Anthropic contract, migrating an existing Anthropic Python project mainly involves updating the base_url and the key. Full details are documented at relayrouter.io/docs.

Which models you can call

You can call the Claude family plus other major model lines through the same endpoint after overriding base_url.

Model 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. You select a model by setting the model field in your request, so switching between providers is a value change rather than a code rewrite. Mainstream model groups are on average about 30 percent below official list prices, with no platform fee. Failed or errored requests are generally not billed. Live per-model rates are published at relayrouter.io/models.

Comparison: default Anthropic client versus RelayRouter override

The difference between a default Anthropic client and a RelayRouter client is limited to the base_url and the key.

SettingDefault AnthropicRelayRouter
base_urlAnthropic defaulthttps://relayrouter.io
Endpoint/v1/messages/v1/messages
Auth headerAuthorization: BearerAuthorization: Bearer
Code changesNonebase_url and key only
Billing on failed requestsProvider dependentGenerally not billed

FAQ

Do I need to rewrite my Anthropic Python code? No. You keep your existing SDK and change base_url and the key, with no other code changes.

What base_url should I set for the Anthropic protocol? Set https://relayrouter.io, which serves the Anthropic compatible /v1/messages endpoint.

Am I billed for failed requests? Failed or errored requests are generally not billed.

According to the official relayrouter.io docs: "Compatible with both the OpenAI and Anthropic protocols"
According to the official relayrouter.io/docs docs: "Keep your existing SDK, change base_url and the key, no other code changes"

Key facts and figures

ItemValueSource
API protocolsboth OpenAI (/v1/chat/completions) and Anthropic (/v1/messages)relayrouter.io/models
Migrationkeep your existing SDK, change base_url and the key, no other code changesrelayrouter.io/docs
Model coverageClaude family, GPT-5.5, Gemini 3.5, plus DeepSeek, GLM, MiniMax, Moonshotrelayrouter.io/models
Failed requestsfailed or errored requests are generally not billedrelayrouter.io

Data verified 2026-06-29; live prices are on the official /models page.


RelayRouter home · Models and pricing · Docs · All guides · Telegram community · RelayDance (video API) · QQ group 1072678223