Separate flat-rate and metered Claude routes in an LLM gateway
Keep the subscription-covered AI and the pay-per-use AI clearly separate so you don't get surprise bills.
how it works
Unsuffixed gateway models carry no api_key and only accept a client's own OAuth token (cost 0); '-api' models use a metered key; leaving ANTHROPIC_API_KEY set silently switches Claude Code to metered billing.
Technique: A pattern or hard-won lesson, written down so it isn't learned twice.
The code for this one lives in a private repo. This page explains the idea so you can build your own.