Add to Cursor / Claude Code

Route AI coding assistants through Routero to bring developer tool usage under the same governance, cost tracking, and policy enforcement as your production applications.

Why route your coding assistant through Routero?

  • Cost visibility — who’s using what models, at what cost, across which teams
  • Policy enforcement — block models or providers that aren’t approved for source code
  • Audit trail — immutable log of all code-assistant calls (useful for IP and data-handling policies)
  • Budget limits — cap individual developer or team spend on AI tooling
  • Provider fallback — coding assistant stays available even if one provider is rate-limited

Cursor

  1. Open Cursor Settings → Models → OpenAI API Key section.
  2. Paste your Routero virtual key as the API key.
  3. Set the Base URL to https://api.routero.ai/v1.
  4. Choose your models — any model string Routero supports works, including smart/balanced.

Cursor will route all its LLM calls through Routero from this point on.

Claude Code (Anthropic CLI)

Set the environment variables before invoking claude:

export ANTHROPIC_API_KEY="YOUR_ROUTERO_KEY"
export ANTHROPIC_BASE_URL="https://api.routero.ai/v1"
claude

Or add to your shell profile (~/.zshrc, ~/.bashrc):

export ANTHROPIC_API_KEY="YOUR_ROUTERO_KEY"
export ANTHROPIC_BASE_URL="https://api.routero.ai/v1"

Claude Code still uses the ANTHROPIC_API_KEY variable name, but with ANTHROPIC_BASE_URL pointing at Routero, all calls route through the gateway. The key you supply should be your Routero virtual key, not an Anthropic key directly.

Generate a per-developer key

Give each developer their own scoped virtual key so spend is attributed individually:

curl -X POST https://api.routero.ai/key/generate \
  -H "Authorization: Bearer $ADMIN_KEY" \
  -d '{
    "team_id": "engineering",
    "models": ["smart/balanced", "anthropic/claude-sonnet-4-6-20250514"],
    "max_budget": 50,
    "budget_duration": "1mo",
    "key_alias": "dev-alice-cursor",
    "metadata": {"developer": "alice@company.com"}
  }'

Per-developer spend appears in the team dashboard alongside production API spend.