The fifth annual Django Developers Survey, prepared by the Django Software Foundation in collaboration with PyCharm, provides a picture of trends in the framework’s community in 2026. The report was based on responses from approximately 3,500 developers in more than 40 countries, ranging from students at the beginning of their journeys to people with decades of experience. The most notable conclusion is that developers are changing tools and working methods much faster than they are changing the framework itself.
PostgreSQL maintained its position as the preferred database for between 76% and 79% of participants throughout five consecutive years, while the Django template engine remained in use by approximately 80% of them. Several months after the release of Django 6.0, 43% said they were already using it, while approximately half of developers move up with each stable release. These figures do not mean that Django is not changing; rather, they indicate that its core architecture provides a stable foundation for experimenting with new tools around it.
AI Is Present, but Its Usage Pattern Has Not Been Settled
AI has become a routine part of developing Django applications. Only 10% of participants said they do not regularly use AI coding tools, while 58% of users use them daily and 27% several times a week. However, the survey does not show a dominant tool or interface; usage is distributed roughly among the browser, the development environment, and the command line.
Claude Code led at 35%, followed by ChatGPT at 33% and GitHub Copilot at 23%. Nevertheless, the most common use remains conversation and asking for advice; 56% of AI users said they use it only in this way. In addition, 59% ask the tool to generate code and then apply the changes themselves, and 44% allow it to modify files or run commands on request, compared with only 27% who use it to complete multi-step tasks independently.
In practice, this describes the agent moving inside the development environment rather than replacing it. The development environment remains the place where developers understand the codebase, review changes, and decide what enters the project. This caution is also reflected in learning: official Django documentation leads learning sources at 67%, but AI tools came second at 51%, ahead of YouTube, reading source code, and Stack Overflow.
Python Tools Are Moving Toward Greater Consolidation
Relatively recent tools for environment management and code quality are standing out. uv, released in February 2024, is used by 43% of participants, ranking after venv at 63% and ahead of Docker at 31%. In linting and formatting, Ruff has become the most-used tool at 43%, ahead of development-environment checks at 27%, Black at 25%, pre-commit at 20%, and Flake8 at 17%.
This does not indicate the disappearance of older tools, but rather that newer tools have begun covering functions that previously required a separate set of tools. For teams, this may mean less fragmented setup paths, but the survey alone does not prove that moving to these tools reduced complexity or maintenance costs.
Types and Testing: Accelerating Adoption and Less Consensus
Type hints have become common practice: 57% of developers use them, and another 26% plan to adopt them. The type checker, however, has not achieved comparable consensus. Among users of type hints, 40% rely on the checker built into their development environment, followed by mypy at 32%, Ruff at 29%, and Pyright/Pylance at 22%. Astral’s ty appeared at 12% and Meta’s Pyrefly at 4%.
This issue is becoming more important as agents take on a larger role in writing code. pytest is used by 45% of participants, unittest by 43%, and pytest-django by 34%. In addition, 51% use GitHub Actions and 26% use GitLab CI/CD. These tools provide a foundation for a cycle in which an agent can modify code, run tests, and respond to errors before presenting the result to the developer. However, 19% of participants do not write automated tests at all, meaning that delegating more work to automated systems is not accompanied by an equivalent verification barrier for everyone.
Two Paths for Building Applications
Seventy-two percent of participants use server-rendered templates, while 53% use Django for API-only applications, and 46% use it behind a single-page application or a separate JavaScript interface. When identifying the primary approach, half of developers build server-rendered applications, while 44% primarily rely on Django for APIs or separate frontends.
JavaScript figures illustrate this split: React’s share remained nearly stable, rising from 37% in 2021 to 38% in the current survey. By contrast, jQuery declined from 37% to 23% and Vue from 28% to 17%, while htmx rose from just 5% to 34%. According to the report’s interpretation, htmx is not taking React’s share; instead, it is revitalizing the server-rendered application segment where jQuery had previously been strongly present.
Why Does This Trend Matter?
The survey results show that Django’s current value is not tied to being the newest framework, but to its ability to accommodate major changes in tools and workflows without requiring a complete rebuild of the application. A team can try a new agent, adopt uv and Ruff, add htmx, or adopt a different type checker while retaining models, the admin panel, and authentication as a familiar foundation.
The conservative tendency also appears in deployment: 54% of participants ship their applications as monoliths, and 44% host their infrastructure themselves. In addition, 33% use Django’s asynchronous features, while 40% plan to adopt them, indicating gradual optional evolution rather than a comprehensive rewrite.
The editorial conclusion here is not that AI will replace the development environment or that one tool will settle the Python tooling race. The data demonstrates widespread use and experimentation, but it does not prove production superiority or improved quality. The most important practical question for teams is how ready their tests and CI/CD pipelines are to verify changes produced by agents, particularly among the share of developers who do not use automated tests.