ENTERPILOT · 2026-04-25 · notable
GoModel — Lightweight AI Gateway in Go, LiteLLM Alternative
Open-source AI gateway in Go: unified OpenAI-compatible endpoint for 10+ providers (OpenAI, Anthropic, Gemini, Groq, xAI, Ollama). 680 GitHub stars, 214 HN Show HN points. Caching, cost tracking, guardrails, and admin dashboard.
A Go-based AI gateway that exposes a single OpenAI-compatible endpoint across 10+ LLM providers, with caching, cost tracking, and an admin dashboard.
What is it?
GoModel is an open-source AI proxy server written in Go. Deploy it in front of your LLM providers (OpenAI, Anthropic, Gemini, Groq, xAI, Azure OpenAI, Oracle, Ollama, vLLM, and more) and point all your apps at a single OpenAI-compatible endpoint. It handles routing, semantic caching, cost attribution, audit logging, and guardrails without touching application code.
How does it work?
GoModel accepts standard OpenAI-format requests, looks up the model alias or direct provider config, forwards the request, and returns a response. For caching it uses exact-match deduplication. Per-user cost attribution is driven by an X-GoModel-User-Path request header, letting you slice usage by team or feature. Guardrails and audit logging are scoped per-provider or per-user-path through a declarative config. Runs as a Docker container or binary with PostgreSQL/MongoDB/Redis backends.
Why does it matter?
LiteLLM is the dominant player in this space but is written in Python, adding latency and memory overhead for Go services. GoModel's native Go implementation stays lightweight for high-throughput applications. 214 HN points on Show HN signals real practitioner interest.
Who is it for?
Backend developers using multiple LLM providers who want centralized routing without Python dependencies
Try it
docker run -p 8080:8080 enterpilot/gomodel