Browse docs
--- title: "Claude Code" description: "Step-by-step guide for routing Claude Code through Aurora with Anthropic passthrough." icon: "terminal" --- Aurora can sit in front of Claude Code so every request goes through your own gateway first. Flow: Claude Code -> Aurora -> Anthropic ## Before you start - Install Claude Code on your machine. - Choose an Aurora master key, for example change-me. - Make sure Aurora has an Anthropic upstream credential. <Note> Claude Code can be routed through Aurora whether or not you personally use a Claude Code subscription. For gateway mode, Claude Code talks to Aurora with ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN. Aurora still needs its own ANTHROPIC_API_KEY to reach Anthropic upstream. </Note> ## How to get ANTHROPIC_API_KEY 1. Open the Claude Console and sign in to your API account. 2. Go to account settings in Console, then create an API key. 3. Copy the key once and set it for Aurora as ANTHROPIC_API_KEY. Anthropic's API docs state that API keys are created in Console account settings. <Warning> Claude paid plans (Pro, Max, Team, Enterprise) and Claude API Console billing are separate. API key usage is billed as API usage. </Warning> ## 1. Run Aurora Start Aurora with a master key and an Anthropic provider key: ``bash docker run --rm -p 8080:8080 \ -e AURORA_MASTER_KEY="change-me" \ -e ANTHROPIC_API_KEY="$ANTHROPIC_API_KEY" \ aurorallm/aurora ` ## 2. Confirm Anthropic passthrough with curl Check that the Anthropic passthrough models endpoint responds: `bash curl -s http://your-aurora-host/p/anthropic/v1/models \ -H "Authorization: Bearer change-me" ` Then send one small test message: `bash curl -s http://your-aurora-host/p/anthropic/v1/messages \ -H "Authorization: Bearer change-me" \ -H "Content-Type: application/json" \ -H "anthropic-version: 2023-06-01" \ -d '{ "model": "claude-3-haiku-20240307", "max_tokens": 16, "messages": [ { "role": "user", "content": "Reply with exactly ok" } ] }' ` If the gateway is wired correctly, the response will contain ok. ## 3. Configure Claude Code to use Aurora Point Claude Code at Aurora's Anthropic passthrough: `bash export ANTHROPIC_BASE_URL=http://your-aurora-host/p/anthropic export ANTHROPIC_AUTH_TOKEN=change-me export CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1 ` Short Claude Code doc summary: for gateway mode, set ANTHROPIC_BASE_URL to your gateway URL and ANTHROPIC_AUTH_TOKEN to your gateway token, then run Claude Code normally. See the official guide: Claude Code LLM gateway docs. If Aurora is not running on your local machine, replace your-aurora-host with the correct host and port. ## 4. Run a Claude Code test prompt `bash claude -p --output-format text --model claude-3-haiku-20240307 \ 'Reply with exactly ok and no punctuation.' ` The validated result was: `text ok ` ## 5. Check the traffic in Aurora Open the Aurora dashboard audit logs: http://your-aurora-host/admin/dashboard/audit This is the easiest place to confirm that Claude Code is reaching Aurora and to inspect the full request and response trail. From the same dashboard, you can keep following your Aurora traffic and usage. ## References - Anthropic Claude Code gateway docs: LLM gateway - Anthropic Claude Code settings: Settings - Claude Help: Managing API key environment variables in Claude Code - Claude Help: Paid Claude plans vs API Console billing - Claude API docs: API overview (keys from Console account settings) ## Validated on March 10, 2026 This guide was validated against: - a local Aurora instance on http://your-aurora-host - an Aurora branch exposing /p/{provider}/v1/... - Claude Code 2.1.72 Local validation confirmed: - GET /p/anthropic/v1/models returned 200 OK - POST /p/anthropic/v1/messages returned 200 OK - claude returned ok` when pointed at Aurora