AIHelms

Why we included this project

AIIHelms suits a company that already wires several providers into one backend and ends up juggling separate API keys, price sheets, and usage dashboards per vendor. The gateway itself routes to OpenAI, Azure, Anthropic, DeepSeek, and self-hosted vLLM or SGLang endpoints; the value sits in the layer wrapped around it. Every call carries token-level accounting, internal and external price tracks are billed separately so you can charge departments an internal rate, and budgets work per person, per project, and per model with soft warnings plus hard cutoffs. It also gives each employee a single API key for all models and runs a central registry for approving and distributing MCP servers and skills. If the nagging question is where the spend went and who is allowed to call what, this is a practical governance layer to examine.