Hi, thanks for the proxy. I’m writing a short checklist for teams that move from a personal Claude Code proxy to a shared gateway/proxy setup, and this repo is directly relevant because users may copy its configuration into team workflows.
Would you consider a docs note covering production/team boundaries?
Suggested content:
- Do not ask users to paste upstream provider master keys into public issues, comments, or shared docs.
- If a proxy/gateway is shared by a team, prefer client-facing gateway keys that can be rotated and revoked separately from upstream provider keys.
- Treat OpenAI/OpenRouter/Ollama/Claude compatibility as verified behavior per provider and feature, not a blanket guarantee.
- Separately validate streaming, tool calling/function calling, thinking/reasoning blocks, and model-specific capabilities before using it in agent workflows.
- Make logging expectations explicit: request metadata and troubleshooting logs are useful, but prompt/content retention and visibility need a privacy boundary.
This is a governance/safety docs suggestion, not a request to bypass provider terms, unlock unlimited usage, or share personal subscriptions. Happy to draft a concise PR if useful.
Hi, thanks for the proxy. I’m writing a short checklist for teams that move from a personal Claude Code proxy to a shared gateway/proxy setup, and this repo is directly relevant because users may copy its configuration into team workflows.
Would you consider a docs note covering production/team boundaries?
Suggested content:
This is a governance/safety docs suggestion, not a request to bypass provider terms, unlock unlimited usage, or share personal subscriptions. Happy to draft a concise PR if useful.