Start with the 50-second overview above: how a high AI bill becomes a tuned proxy, step by step, with an illustrative example (not a customer result). It has captions and plays without sound.
This is a recorded walkthrough, not a live event. It follows one admin setting up a company from nothing in the Proxen console, in the order a real setup happens. Ask for the recording and we will send the link, or book a call and we will do the same steps on your account.
What the walkthrough covers
- Connecting accounts. Adding the company's own Anthropic, OpenAI and Google keys. Where they are stored, who can see them (nobody on the team), and how to rotate one.
- Teams and model lists. Creating an engineering team and a content team, giving each a different list of models, and what a person sees when they ask for a model that is not on their list.
- Issuing keys. One key per person, one per service, and the safe way to hand a key to its owner.
- Budgets and caps. Setting a monthly budget per person inside a team budget, the alert at 80 percent, the hard stop at 100, and raising a budget mid-month.
- The spend view. Live spend by person, team and model, and the saving from each trimming, caching, routing and batch rule. What is logged (who, model, tokens, cost) and what is not (prompt bodies, unless retention is switched on for a team).
- Revoking a key. The one click that ends a leaver's access without touching anyone else.
Who it is for
The founder or ops lead who will administer Proxen, and anyone in IT or security who wants to see where the provider keys live before signing off.