Can I use LiteLLM with my existing OpenAI SDK code?

Yes. Just change the `base_url` to your LiteLLM endpoint and the `api_key` to your master key. Everything else stays identical.

Does LiteLLM add latency?

Minimal: a few milliseconds of proxy overhead. For most applications this is unnoticeable.

Can I create separate API keys for different applications?

Yes. LiteLLM supports virtual keys with individual rate limits and spending caps. Create them in the dashboard or via the API.

LiteLLM on a VPS: One API Endpoint for Every LLM Provider

What LiteLLM Does

Single OpenAI-compatible endpoint for 100+ providers
Route requests between providers based on model name
Track costs per model, per team, per API key
Set spending limits and rate limits
Fall back to backup providers if primary fails
Load balance across multiple deployments

What You Need

A VPS with at least 2 GB RAM
Docker installed
API keys for whichever providers you want to use

Step 1: Create the Configuration File

mkdir ~/litellm && cd ~/litellm
nano config.yaml

model_list:
  - model_name: gpt-4o
    litellm_params:
      model: openai/gpt-4o
      api_key: os.environ/OPENAI_API_KEY

  - model_name: claude-sonnet
    litellm_params:
      model: anthropic/claude-sonnet-4-6
      api_key: os.environ/ANTHROPIC_API_KEY

  - model_name: local-llama
    litellm_params:
      model: ollama/llama3.1:8b
      api_base: http://host.docker.internal:11434

general_settings:
  master_key: your-master-key-here

Step 2: Deploy with Docker

nano docker-compose.yml

version: '3.8'
services:
  litellm:
    image: ghcr.io/berriai/litellm:main-latest
    restart: always
    ports:
      - "4000:4000"
    volumes:
      - ./config.yaml:/app/config.yaml
    environment:
      - OPENAI_API_KEY=${OPENAI_API_KEY}
      - ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY}
      - LITELLM_MASTER_KEY=${LITELLM_MASTER_KEY}
    command: --config /app/config.yaml --port 4000
    extra_hosts:
      - "host.docker.internal:host-gateway"

Create a .env file:

OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
LITELLM_MASTER_KEY=your-master-key-here

docker compose up -d

Step 3: Test the Endpoint

curl http://localhost:4000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-master-key-here" \
  -d '{
    "model": "local-llama",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Switch models by changing the model name, same endpoint, same format.

Step 4: Expose with Nginx and HTTPS

apt install -y nginx certbot python3-certbot-nginx
nano /etc/nginx/sites-available/litellm

server {
    server_name llm.yourdomain.com;

    location / {
        proxy_pass http://localhost:4000;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}

ln -s /etc/nginx/sites-available/litellm /etc/nginx/sites-enabled/
certbot --nginx -d llm.yourdomain.com
nginx -t && systemctl reload nginx

The Cost Tracking Dashboard

LiteLLM ships with a built-in dashboard at /ui. Log in with your master key to see spend per model, per virtual key, and per team. Set budget limits to prevent runaway costs.

What it costs

LiteLLM is a lightweight proxy, so it barely uses resources itself; the heavy lifting stays with whichever provider you route to. The Starter plan (1 vCPU, 2 GB) matches the 2 GB minimum above and runs it fine as a single gateway for your team. NoctHost bills hourly from a prepaid balance you top up with crypto (no KYC) or a card.

LiteLLM on a VPS: One API Endpoint for Every LLM Provider

What LiteLLM Does

What You Need

Step 1: Create the Configuration File

Step 2: Deploy with Docker

Step 3: Test the Endpoint

Step 4: Expose with Nginx and HTTPS

The Cost Tracking Dashboard

What it costs

Spin one up in about a minute

Frequently asked

Keep reading

Run Ollama on a VPS

Open WebUI on a VPS

Best local LLM models for a VPS