Productive in
5 minutes.

Step 01

Create account & workspace

Register at mycelis.io/register. After login, you are guided through workspace creation. A workspace is your isolated area for deployments, agents, and API keys.

Step 02

Select & start deployment

Choose between three deployment options:

Navigate to Model Marketplace → Select model → Start deployment. A GPU instance is ready in under 60 seconds.

Step 03

Create API key & use API

Under API Keys to create a Personal Access Token (PAT). In your existing code, only replace base_url and api_key :

# OpenAI-compatible — no SDK switch required
from openai import OpenAI

client = OpenAI(
    base_url="https://api.mycelis.io/proxy/v1",
    api_key="dein-pat-key"   # from Dashboard → API Keys
)

response = client.chat.completions.create(
    model="llama-3.1-8b-instruct",   # deployment slug
    messages=[{
        "role": "user",
        "content": "Explain quantum computing in 3 sentences."
    }]
)

print(response.choices[0].message.content)
# Fully private. No data leaves your infrastructure.

Step 04 — Optional

Configure agent

Under Agents create a VirtualModel. Link a deployment, set a system prompt, add knowledge bases (RAG), and enable Smart Routing for automatic cost control.

After that, use your agent slug instead of the direct deployment slug: