GoModel

Why we included this project

GoModel acts as a single front door for model calls, exposing one OpenAI- and Anthropic-compatible API so your application only ever talks to that endpoint. It can map a stable model name across several upstreams, fail over to a backup provider when one goes down, and reroute to a cheaper option when spending matters. The bundled dashboard turns raw usage into per-request cost estimates, budgets, and rate limits per key or team, which most teams end up building by hand. It ships as a single Go binary or Docker container and works with the official OpenAI and Anthropic SDKs as-is, since you just point the base URL at it. Teams used to LiteLLM-style gateways who want something lighter and self-hosted will find the workflow familiar.