Set up a private LLM gateway for your team with keys and limits

Your team uses ChatGPT, Claude and a couple of other models. Some people have their own keys, others share one that was pasted into a group chat months ago. Nobody knows who spent what, and client data goes into prompts that nobody checks.

I'll put one gateway in front of your LLM providers. Your apps and staff point at a single OpenAI-compatible URL, and the gateway handles keys, limits and logging.

What you get:
- One endpoint that works with any tool built for the OpenAI API
- Routing to up to 3 providers (OpenAI, Anthropic, Google or a self-hosted model)
- A personal key for each person or app, which you can revoke in one click
- A monthly spending limit per key, after which requests are blocked
- Rules that block or mask emails, card numbers or words you choose before a request leaves your network
- Request logs with user, model, tokens and cost, kept for as long as you set
- A small admin panel to manage keys, limits and rules
- Deployment with Docker on your server, plus a one-page setup guide

I've built and run a gateway like this before: an OpenAI-compatible layer in front of several LLM providers, with client keys, rules, logs and an admin panel. Six years of commercial full-stack development behind it.

Before you order, tell me which providers you use, roughly how many people or apps will connect, and where you want it hosted. You pay the providers and the server directly. Single sign-on with Google or Microsoft accounts is available as an add-on.