Integration Guides

These guides show how to point popular clients and agents at the ru-llm gateway. ru-llm exposes an OpenAI-compatible surface (/v1/chat/completions) and an Anthropic-compatible surface (/v1/messages) on a single base URL, so most tools work after changing two settings: the base URL and the API key.

What you need

• Base URL — https://api.ru-llm.relay2.xyz (replace with your gateway URL if self-hosted).
  • OpenAI-style clients point at https://api.ru-llm.relay2.xyz/v1.
  • Anthropic-style clients point at https://api.ru-llm.relay2.xyz.
• API key — an sk-ru-… key from the console → API Keys. It is accepted as Authorization: Bearer sk-ru-… or x-api-key: sk-ru-… on both surfaces.
• A model ID — see Models & Pricing. Chat examples below use gpt-5.5, claude-sonnet-4-6, and claude-opus-4-8.

First request

OpenAI-compatible (chat completions)

curl https://api.ru-llm.relay2.xyz/v1/chat/completions \
  -H "Authorization: Bearer sk-ru-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [{"role": "user", "content": "Hello, world!"}]
  }'

Response:

{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "model": "gpt-5.5",
  "choices": [
    {"index": 0, "message": {"role": "assistant", "content": "Hello! How can I help you today?"}, "finish_reason": "stop"}
  ],
  "usage": {"prompt_tokens": 9, "completion_tokens": 10, "total_tokens": 19}
}

Anthropic-compatible (messages)

curl https://api.ru-llm.relay2.xyz/v1/messages \
  -H "x-api-key: sk-ru-YOUR_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 256,
    "messages": [{"role": "user", "content": "Hello, world!"}]
  }'

Cross-dialect

The gateway translates between dialects. You can send a Claude model ID to /v1/chat/completions, or a GPT/Gemini model ID to /v1/messages — the proxy translates both the request and the response, including streaming. So a client that only speaks one dialect can still reach every chat model.

Pick your client

Claude Code

Point Anthropic’s CLI/IDE agent at the gateway with two env vars. Open guide →

Codex

Add a custom OpenAI provider in config.toml. Open guide →

Image generation

Generate images with gpt-image-2 via /v1/images/generations. Open guide →

Video generation

Render video with Seedance via the async video endpoint. Open guide →

Generic OpenAI / Anthropic SDKs

Any client built on the official OpenAI or Anthropic SDKs works by overriding the base URL and key — see the OpenAI-compatible and Anthropic-compatible references for Python and TypeScript snippets.