- Role
Virtual keys
Loading keys…
No virtual keys associated with this account.
| Name | Key | Team | Allowed models | Spend | Budget | Expires | Status | Actions |
|---|---|---|---|---|---|---|---|---|
Teams
Loading teams…
No teams with an invite link found on this server.
Playground
Send a message to test a virtual key against the proxy.
Help
Calling the proxy from Python
Use any virtual key from the Virtual keys table as the API key of a standard OpenAI client, pointing it at this proxy server (). The proxy forwards requests to the real model providers and tracks the spend of each key.
The proxy speaks the standard OpenAI HTTP API, so any OpenAI-compatible client library, in any language, can be used the same way — provided it is pointed at the proxy URL () with a virtual key.
How budgets work
- Every virtual key has a spend, counted in US dollars by the proxy on each request, and an optional budget (USD). Once a key's spend reaches its budget, the proxy rejects further requests made with that key. Keys without a budget show No limit in the table.
- A team budget caps the combined spend of all the keys belonging to a team — once the team total is reached, every key of the team stops working, even keys with remaining individual budget.
- A team member budget is the default budget automatically given to the virtual key created when someone joins the team through an invite link. Team admins set it when creating the team; it is shown as Member budget in the Teams table and as Default key budget in the invitation page.
- The Spend and Budget columns of the Virtual keys table (and the Spend column of the Teams table) always show the current values tracked by the proxy. Regenerating a key keeps its budget and accumulated spend.
LiteLLM
Sign in to the proxy server