Using LangChain ChatAnthropic with RelayRouter's Anthropic protocol endpoint

To use LangChain's ChatAnthropic with RelayRouter, keep the ChatAnthropic class you already have, point its base URL at RelayRouter's Anthropic compatible endpoint (POST /v1/messages, base https://relayrouter.io), and swap in your RelayRouter API key. RelayRouter is «Compatible with both the OpenAI and Anthropic protocols» (relayrouter.io), so no other code changes are required. Authenticate with Authorization: Bearer YOUR_API_KEY and you can call the Claude family through the same interface.

How the Anthropic endpoint works

RelayRouter exposes an Anthropic compatible route at POST /v1/messages with the base URL https://relayrouter.io, which LangChain's ChatAnthropic can target directly. According to the official relayrouter.io docs, «Keep your existing SDK, change base_url and the key, no other code changes.» That means the same ChatAnthropic invocation, message roles and streaming behaviour you use today continue to function; you only redirect the base URL and provide the RelayRouter key. Authentication uses the header Authorization: Bearer YOUR_API_KEY, and keys are created at https://relayrouter.io/dashboard. Because the gateway speaks both the OpenAI (/v1/chat/completions) and Anthropic (/v1/messages) protocols, you can standardise on ChatAnthropic while retaining access to a wider model set. See details at relayrouter.io/docs.

Configuration steps

Migration to RelayRouter with ChatAnthropic takes three steps and no structural code changes.

1. Set the base URL: configure ChatAnthropic to use the Anthropic compatible base https://relayrouter.io, which routes to POST /v1/messages.
2. Set the key: create an API key at https://relayrouter.io/dashboard and pass it so requests send Authorization: Bearer YOUR_API_KEY.
3. Choose a model: select a supported model such as a Claude family model, then run your existing invocation.

Because RelayRouter follows the rule to «Keep your existing SDK, change base_url and the key, no other code changes» (relayrouter.io/docs), your prompt logic, message formatting and streaming code stay intact. Live per-model rates are published at relayrouter.io/models.

Model coverage and protocols

Through the Anthropic endpoint you can reach the Claude family plus other model groups routed by the gateway. Model coverage includes the Claude line (for example claude-opus-4-8 and claude-fable-5), gpt-5.5, Gemini 3.5 (gemini-3.5-flash), and additional providers DeepSeek, GLM, MiniMax and Moonshot. The table below summarises the two request protocols relevant to SDK migration.

Streaming is supported on these routes. For full model coverage and current identifiers, see relayrouter.io/models.

Pricing and billing behaviour

Mainstream model groups on RelayRouter are on average about 30 percent below official list prices, with no platform fee. Payments are handled by Stripe card, and API keys are managed at https://relayrouter.io/dashboard. Failed or errored requests are generally not billed, so ChatAnthropic calls that return an error do not add cost. Because pricing is set per model, always confirm the current rate for a Claude family model (for example claude-opus-4-8) at the live rates page before running production workloads. This billing model applies uniformly whether you call through the Anthropic compatible route (POST /v1/messages) or the OpenAI compatible route (POST /v1/chat/completions). Live per-model rates are listed at relayrouter.io/models.

Frequently asked questions

Do I need to rewrite my LangChain ChatAnthropic code? No. You keep your existing SDK, change the base URL to https://relayrouter.io and set the RelayRouter key; no other code changes are required.

Which base URL and endpoint does ChatAnthropic use? The Anthropic compatible endpoint is POST /v1/messages with base https://relayrouter.io, authenticated via Authorization: Bearer YOUR_API_KEY.

Am I charged for failed requests? No. Failed or errored requests are generally not billed. Confirm current per-model rates at relayrouter.io/models.