Un réseau d'inférence distribué propulsé par des GPU grand public et Tailscale

export OPENAI_API_KEY=ic_xxxxxxxxxxxxxxxxxxxx
export OPENAI_BASE_URL=https://api.inference.club/v1

curl $OPENAI_BASE_URL/chat/completions \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3.6-27b",
    "messages": [
      {"role": "user", "content": "explain MoE in one sentence"}
    ]
  }'

from openai import OpenAI

client = OpenAI(
    api_key="ic_xxxxxxxxxxxxxxxxxxxx",
    base_url="https://api.inference.club/v1",
)

resp = client.chat.completions.create(
    model="qwen/qwen3.6-27b",
    messages=[
        {"role": "user", "content": "explain MoE in one sentence"},
    ],
)
print(resp.choices[0].message.content)

import OpenAI from "openai"

const client = new OpenAI({
  apiKey: process.env.INFERENCE_CLUB_KEY,
  baseURL: "https://api.inference.club/v1",
})

const resp = await client.chat.completions.create({
  model: "qwen/qwen3.6-27b",
  messages: [
    { role: "user", content: "explain MoE in one sentence" },
  ],
})

console.log(resp.choices[0].message.content)

# Already running vLLM, llama.cpp, or Ollama on your GPU?
# Point the agent at it and join the network.

export INFERENCE_CLUB_API_KEY=ic_xxxxxxxxxxxxxxxxxxxx
export OPENAI_BASE_URL=http://localhost:8000/v1
export OPENAI_API_KEY=local-key  # whatever your local server expects

docker run --rm -d --name club-agent --network host \
  -e INFERENCE_CLUB_API_KEY \
  -e OPENAI_BASE_URL \
  -e OPENAI_API_KEY \
  ghcr.io/inference-club/inference-club-agent:latest

# Or just run the static binary — no Docker required.

export INFERENCE_CLUB_API_KEY=ic_xxxxxxxxxxxxxxxxxxxx
export OPENAI_BASE_URL=http://localhost:8000/v1
export OPENAI_API_KEY=local-key

./inference-club-agent

# The agent registers, joins the inference.club tailnet,
# advertises the models your local server is serving,
# and starts taking requests. That's it.