# Use Claude Code with AnyRouter

> Route Claude Code through AnyRouter for unified billing, audit logs, and a single API key that unlocks every Anthropic model. Drop-in via two environment variables.


# Use Claude Code with AnyRouter

Route Claude Code through AnyRouter to get unified billing, request audit logs, and a single API key for every Anthropic model — all by changing two environment variables.

## Why

- **Drop-in.** Zero code changes. Set two env vars and Claude Code works.
- **One key.** A single AnyRouter API key unlocks every Anthropic model and more.
- **Audit logs.** Every request is logged with cost, latency, and token usage.

## Step 1 — Create an AnyRouter API key

Head to your [dashboard](/dashboard/keys) and create a new API key. It will be prefixed with `ar-`.

## Step 2 — Point Claude Code at AnyRouter

Claude Code reads `ANTHROPIC_BASE_URL` and `ANTHROPIC_API_KEY` from the environment. Set them to your AnyRouter base URL and API key:

```bash
export ANTHROPIC_BASE_URL="https://anyrouter.dev/api"
export ANTHROPIC_API_KEY="ar-your-anyrouter-key"
```

Drop these two lines into `~/.bashrc` or `~/.zshrc` so they persist across shells.

## Step 3 — Run Claude Code

That's it. Every Claude Code request now routes through AnyRouter. Requests are billed against your AnyRouter credits and recorded in your audit logs.

```bash
# Run Claude Code against AnyRouter
claude "Explain the architecture of this project"

# Specify a model explicitly
claude --model anthropic/claude-sonnet-4.6 "Refactor this function"

# Use the faster/cheaper Haiku model
claude --model anthropic/claude-haiku-4.5 "Summarize this file"
```

## Using the Messages API directly

Any client that targets Anthropic's Messages API works. Send a POST to `/api/v1/messages` with the `x-api-key` header.

```bash
curl -sS https://anyrouter.dev/api/v1/messages \
  -H "x-api-key: ar-your-anyrouter-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-haiku-4.5",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Say hi in 5 words"}
    ]
  }'
```

Streaming works the same way — just add `"stream": true`:

```bash
curl -sSN https://anyrouter.dev/api/v1/messages \
  -H "x-api-key: ar-your-anyrouter-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-4.6",
    "max_tokens": 1024,
    "stream": true,
    "messages": [{"role": "user", "content": "Count to 5"}]
  }'
```

## Supported models

The `/v1/messages` endpoint currently supports all `anthropic/*` models as a native passthrough. For non-Anthropic models (OpenAI, Gemini, Llama, Qwen, DeepSeek, etc.), use the `/v1/chat/completions` OpenAI-compatible endpoint instead.

- `anthropic/claude-sonnet-4.6` — flagship, 1M context
- `anthropic/claude-haiku-4.5` — fastest and cheapest
- `anthropic/claude-opus-4.6` — highest capability

:::note
Cross-provider routing via the Messages API (e.g. using `openai/gpt-4o` through `/v1/messages`) is coming in Phase 2. For now, pair Claude Code with Anthropic models and use `/v1/chat/completions` for everything else.
:::