How do you set LLM budgets per team?
retained text
We're about 25 people. Engineering, support, and marketing all use llms through the API, and right now it's one shared key with no visibility. Last month someone's test script burned a painful chunk of the budget over a weekend. ideal world: \- A separate key per team or project, each with a hard limit \- A live view of who's spending what, on which models \- To avoid becoming the person who approves every request I'm leaning toward a gateway instead of building internal tooling. llmapi.ai has per-key limits and usage analytics, and LiteLLM can do something similar if you self-host it. Has anyone rolled out either one company-wide? What's worked without adding a pile of process?
document fetched 2026-09-30 00:26 · canonical https://reddit.com/r/devops/comments/1wt1llw/how_do_you_set_llm_budgets_per_team