djgent 0.4.0

Django AI Agent Framework - Build AI agents with LangChain integration in Django

Djgent

Django AI Agent Framework for building AI agents with LangChain integration in Django applications.

Installation

pip install djgent

Or install with uv:

uv add djgent

Basic Setup

Add djgent to your Django project:

INSTALLED_APPS = [
    # ...
    "djgent",
]

Configure your LLM provider:

DJGENT = {
    "DEFAULT_LLM": "openai:gpt-4o-mini",
    "API_KEYS": {
        "OPENAI": "your-openai-api-key",
    },
}

Run migrations when using database-backed memory or the built-in chat UI:

python manage.py migrate

Default Chat View

Djgent includes a default chat UI you can mount in your project.

Add the chat app:

INSTALLED_APPS = [
    # ...
    "djgent",
    "djgent.chat",
]

Configure the chat:

DJGENT = {
    "DEFAULT_LLM": "openai:gpt-4o-mini",
    "API_KEYS": {
        "OPENAI": "your-openai-api-key",
    },
    "CHAT_UI": {
        "TITLE": "AI Assistant",
        "TOOLS": ["calculator", "datetime"],
        "AUTO_LOAD_TOOLS": True,
        "SYSTEM_PROMPT": "You are a helpful assistant.",
    },
}

Mount the default chat URLs:

from django.urls import include, path

urlpatterns = [
    path("chat/", include("djgent.chat.urls")),
]

Then open /chat/ in your browser.

To add the optional site-wide chat bubble, enable it in settings:

DJGENT = {
    # ...
    "CHAT_UI": {
        "BUBBLE_ENABLED": True,
        "BUBBLE_TITLE": "Ask AI",
    },
}

And render it in your base template:

{% load djgent_chat %}
{% djgent_chat_bubble %}

Create and run an agent:

from djgent import Agent

agent = Agent.create(name="assistant")
response = agent.run("What is Djgent?")
print(response)

Full Documentation

See the full Djgent documentation on GitHub.