Claude Max API Proxy: OpenAI-Compatible Access to Claude Subscriptions

claude-max-api-proxy is a community-maintained npm package, not an official OpenClaw plugin. It turns a Claude Max or Pro subscription into an OpenAI-compatible API endpoint, letting any tool that speaks OpenAI's format talk to your subscription rather than requiring an Anthropic API key.

Warning

This is a technical compatibility layer, not an officially endorsed route. Anthropic has previously restricted some subscription usage outside Claude Code; check Anthropic's current billing terms before depending on it.

Anthropic's Claude Code documentation refers to claude -p as Agent SDK or programmatic usage. Per Anthropic's June 15, 2026 support update, Claude Agent SDK, claude -p, and third-party app usage all count against the signed-in subscription's usage limits (the earlier announced separate Agent SDK credit plan is on hold). Refer to Anthropic's Agent SDK plan article, the Pro/Max and Team/Enterprise plan articles, and Anthropic provider for OpenClaw's own Claude CLI billing notes.

Why use this

With this proxy, a Claude Max or Pro subscription becomes usable from OpenAI-compatible tools. It is not a flat-rate unlimited option; it follows Claude Code's usage limits. For production, API keys remain the more straightforward billing choice.

How it works

Your App -> claude-max-api-proxy -> Claude Code CLI / claude -p -> Anthropic
     (OpenAI format)                (converts format)              (uses your login)

Each request causes the proxy to launch the Claude Code CLI as a subprocess, translate OpenAI-format chat messages into CLI prompts, and then stream or return the reply in OpenAI format.

Getting started

Install the proxy

You need Node.js 20 or newer and an authenticated Claude Code CLI.

npm install -g claude-max-api-proxy

# Verify Claude CLI is authenticated
claude --version
claude auth login   # if not already authenticated

Start the server

claude-max-api
# Server runs at http://localhost:3456

Test the proxy

curl http://localhost:3456/health
curl http://localhost:3456/v1/models

curl http://localhost:3456/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-opus-4",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Configure OpenClaw

Configure OpenClaw to use the proxy as a custom OpenAI-compatible endpoint:

{
  env: {
    vars: {
      OPENAI_API_KEY: "not-needed",
      OPENAI_BASE_URL: "http://localhost:3456/v1",
    },
  },
  agents: {
    defaults: {
      model: { primary: "openai/claude-opus-4" },
    },
  },
}

Note

The model ids shown below come from the proxy's own catalog, not from OpenClaw's Anthropic model refs. Each id corresponds to a Claude Code CLI model alias (opus, sonnet, haiku), so the actual model changes whenever Anthropic updates that alias in the CLI. Always check the proxy's current README before relying on a particular mapping.

Advanced configuration

Proxy-style OpenAI-compatible notes

OpenClaw's generic custom /v1 OpenAI-compatible route handles this, the same path used for any other self-hosted OpenAI-compatible backend:

• OpenAI-only request shaping is not applied.
• /fast and service_tier only affect direct api.anthropic.com traffic; proxy routes leave service_tier untouched (see Anthropic provider fast mode).
• No Responses store, prompt-cache hints, or OpenAI reasoning-compat payload shaping.
• OpenClaw's OpenAI/Codex attribution headers (originator, version, User-Agent) go out only on native api.openai.com OAuth traffic, not on custom OPENAI_BASE_URL targets such as this proxy.

Auto-start on macOS with LaunchAgent

cat > ~/Library/LaunchAgents/com.claude-max-api.plist << 'EOF'
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
  <key>Label</key>
  <string>com.claude-max-api</string>
  <key>RunAtLoad</key>
  <true/>
  <key>KeepAlive</key>
  <true/>
  <key>ProgramArguments</key>
  <array>
    <string>/usr/local/bin/node</string>
    <string>/usr/local/lib/node_modules/claude-max-api-proxy/dist/server/standalone.js</string>
  </array>
  <key>EnvironmentVariables</key>
  <dict>
    <key>PATH</key>
    <string>/usr/local/bin:/opt/homebrew/bin:~/.local/bin:/usr/bin:/bin</string>
  </dict>
</dict>
</plist>
EOF

launchctl bootstrap gui/$(id -u) ~/Library/LaunchAgents/com.claude-max-api.plist

Notes

• Follows Claude Code's claude -p billing, usage-credit, and rate-limit behavior.
• Listens on 127.0.0.1 only; no data reaches third-party servers beyond the CLI's own call to Anthropic.
• Streaming responses work.
• Auth failures aren't detected at startup; they only show up when a chat request actually executes. If the CLI isn't authenticated, expect the first request to error rather than the server refusing to launch.

Note

For native Anthropic integration with Claude CLI or API keys, see Anthropic provider. For OpenAI/Codex subscriptions, see OpenAI provider.

Related

• Anthropic provider, Native OpenClaw integration with Claude CLI or API keys.

• OpenAI provider, For OpenAI/Codex subscriptions.

• Model selection, Overview of all providers, model refs, and failover behavior.

• Configuration, Full config reference.