Recipe

Django Primer

A focused primer for wiring Meridian into a Django backend. This recipe walks through project layout, installing the SDK, and shipping your first AI-powered view without rewriting your stack. Read top to bottom or jump to the section you need.

1.Project setup

Start from a fresh virtualenv. Meridian targets Django 4.2+ and Python 3.10+. Pin the SDK inrequirements.txtso deploys stay reproducible across staging and prod.

python -m venv .venv
source .venv/bin/activate
pip install django meridian-sdk
django-admin startproject mysite
cd mysite && python manage.py startapp ai

2.Configure the client

Add your Meridian API key to environment variables and build a singleton client in your app config. Avoid instantiating the client per-request — connection reuse keeps p95 latency well under 200ms on the gateway.

Store the key in MERIDIAN_API_KEY
Initialize in AppConfig.ready()
Expose through Django settings

3.Your first view

With the client live, a view becomes a thin wrapper around client.chat(). Stream tokens back with Django's StreamingHttpResponse for interactive UX, or return JSON for backend pipelines.

from django.http import JsonResponse
from meridian import Meridian

client = Meridian()

def summarize(request):
    text = request.GET.get("q", "")
    out = client.chat(
        model="azure/model-router",
        messages=[{"role": "user", "content": text}],
    )
    return JsonResponse({"reply": out.content})