Every model.
Every language.
On your box.

# 1. clone & build — embeds the web UI
git clone https://github.com/vul-os/llmux
cd llmux
make build

# 2. configure a provider (or skip — auto-detected from env)
export OPENAI_API_KEY=sk-...

# 3. run — gateway on :4000, dashboard at /ui
cp llmux.example.json llmux.json
./dist/llmux -config llmux.json# the repo ships a Dockerfile — no local Go toolchain needed
docker build -t llmux .
docker run -p 4000:4000 \
  -e LLMUX_MASTER_KEY=$(openssl rand -hex 32) \
  -e OPENAI_API_KEY=sk-... \
  -v $PWD/llmux.json:/etc/llmux/llmux.json \
  llmux -config /etc/llmux/llmux.json

from openai import OpenAI

client = OpenAI(base_url="http://localhost:4000/v1", api_key="sk-team-a")

resp = client.chat.completions.create(
    model="cheapest",                  # least-cost live route
    messages=[{"role": "user", "content": "hi"}],
)
print(resp.usage)  # includes per-request cost_usdimport OpenAI from "openai";

const client = new OpenAI({ baseURL: "http://localhost:4000/v1" });

const r = await client.chat.completions.create({
  model: "anthropic/claude-3-5-sonnet",  // provider/model prefix
  messages: [{ role: "user", content: "hi" }],
});curl http://localhost:4000/v1/chat/completions \
  -H "Authorization: Bearer sk-team-a" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-*","messages":[{"role":"user","content":"hi"}]}'
# streaming ("stream":true) returns byte-identical OpenAI SSE{
  "providers": [
    { "name": "local",  "base_url": "http://127.0.0.1:11434/v1" },  // on-box: always allowed
    { "name": "openai", "api_key_env": "OPENAI_API_KEY",
      "allow_egress": true }        // explicit off-box opt-in
  ],
  "routes": [
    { "model": "*", "provider": "local", "fallbacks": ["openai"] }
  ]
}