Welcome to the highlights from the fifth annual Django Builders Survey, a collaboration between the Django Software program Basis and PyCharm. This 12 months’s report attracts on responses from almost 3,500 Django builders throughout greater than 40 international locations — from college students of their first 12 months to veterans with a long time of expertise.
In software program, “boring” is a praise. It means a expertise works so reliably you possibly can cease desirous about it. By that normal, Django in 2026 is prospering: PostgreSQL has been the database of selection for 76–79% of respondents for 5 consecutive years, Django’s template engine has held regular at round 80%, and almost half of builders improve with each steady launch — 43% are already on Django 6.0, months after it shipped. The core is boring in the absolute best manner. Mature, however not static.
All the pieces round that core, although, is shifting quick. AI is now an on a regular basis device for many builders, and brokers are starting to maneuver past answering inquiries to modifying information, operating instructions, and finishing bigger duties. Newer instruments are consolidating workflows that when required a number of separate utilities, typing is changing into normal apply, and the boundaries between editors, terminals, automation, and AI are getting blurrier.
Which may be Django’s explicit benefit in 2026: it’s mature sufficient to be reliable, lively sufficient to maintain shifting, and steady sufficient to offer builders room to vary nearly all the things else.
As an open-source framework, Django relies on its group and wishes funding to stay wholesome and safe. The PyCharm group runs an annual fundraiser to assist Django.
Till September 10, get PyCharm at 30% off, and JetBrains will donate 100% of your buy quantity to the Django Software program Basis.
1. AI is mainstream, however no workflow has gained
AI has change into a part of the traditional Django improvement workflow. Solely 10% of respondents stated they recurrently use no AI instruments for coding, whereas 58% of AI customers use them every single day and one other 27% a number of occasions every week.
What stays unsettled is how builders use them. The interfaces are break up nearly evenly throughout the browser, the IDE, and the command line, and no device dominates: Claude Code leads at 35%, with ChatGPT simply behind at 33% and GitHub Copilot at 23%. And for all the eye on brokers, a majority of AI customers — 56% — nonetheless use it purely for chat and recommendation.
2. Dominating AI workflow remains to be supervised
Practically half of AI customers already work with AI by an IDE integration. However regardless of the expansion of coding brokers, the dominant workflow remains to be supervised: 59% have AI generate code after which apply the modifications themselves, 44% let it edit information or run instructions when instructed, and solely 27% use it to autonomously full multi-step duties.
The rising sample is subsequently much less “exchange the IDE with an agent” and extra “convey the agent into the event surroundings.” Builders are adopting AI shortly, however the editor stays their house base for understanding the codebase, reviewing modifications, and deciding what makes it into the challenge.
Nowhere is the shift clearer than in how builders be taught. Django’s official documentation stays the highest useful resource at 67%, however AI instruments at the moment are second at 51% — forward of YouTube, studying supply code, and Stack Overflow.
AI is already routine for writing code, debugging, analysis, and studying. However builders are nonetheless experimenting with the place it belongs: within the browser, contained in the IDE, on the terminal, or more and more, appearing immediately on the codebase.
AI is altering the IDE quicker than it’s changing it.
3. Python tooling is consolidating quick
Two instruments that hardly existed a number of years in the past are already close to the highest of the survey.
uv, launched in February 2024, is already utilized by 43% of respondents for managing Python environments—second solely to venv at 63% and forward of Docker at 31%.
Ruff tells the same story. At 43%, it’s now probably the most broadly used code-quality and formatting device within the survey, forward of IDE inspections at 27%, Black at 25%, pre-commit at 20%, and Flake8 at 17%.
The shift just isn’t that Python’s older instruments have disappeared. It’s that newer instruments more and more cowl jobs that when required a number of separate utilities. The result’s a Python toolchain starting to consolidate round fewer, quicker, extra succesful instruments.
4. Sort hints are successful. The kind checker race is huge open.
Sort hints have gotten the norm in Django improvement: 57% already use them, and one other 26% plan to. What is far much less settled is how builders verify these sorts.
Amongst builders who use kind hints, the most well-liked possibility isn’t a standalone kind checker in any respect: 40% depend on the checker constructed into their IDE. Mypy follows at 32%, Ruff at 29%, and Pyright/Pylance at 22%. Newer entrants are already showing too, with Astral’s ty reaching 12% and Meta’s Pyrefly at 4%.
The apply, in different phrases, is converging quicker than the tooling. Django builders more and more agree that sorts are helpful, however there’s nonetheless no consensus on which device ought to implement them—or whether or not a separate device is important in any respect.
That makes kind checking an attention-grabbing area to look at in 2027: will one of many newer standalone instruments break free from the pack, or will kind checking more and more change into one thing builders merely anticipate their IDE to supply?
5. As AI writes extra code, verification issues extra
AI is shifting past strategies and into the codebase. That makes automated verification extra necessary, not much less.
The survey can’t inform us whether or not AI is driving higher adoption of exams or CI. What it does present is that almost all Django builders have already got the infrastructure brokers want: pytest is utilized by 45% of respondents and unittest by 43%, with pytest-django at 34%.
That testing tradition sits alongside widespread CI/CD adoption — GitHub Actions is now utilized by a majority of respondents at 51%, with GitLab CI/CD at one other 26%. Collectively, these create a pure suggestions loop for agentic improvement: an agent could make a change, run the check suite, reply to failures, and hand the developer a end result that has already handed the challenge’s checks.
Not everybody has that loop in place: 19% of respondents write no automated exams in any respect. As extra code is delegated to brokers, that fifth of builders is working with out the protection web that makes delegation reliable.
The rising agentic workflow might rely as a lot on verification as technology. The extra code we delegate, the extra priceless it turns into to have exams and pipelines that may shortly inform each builders and brokers whether or not a change truly works.
Brokers can generate code. Assessments and pipelines inform them whether or not it really works.
6. One framework, two methods to construct
Two distinct methods of constructing with Django at the moment are firmly mainstream: letting Django render the interface, or utilizing it because the backend for a separate frontend. 72% of respondents use server-rendered templates, whereas 53% use Django for API-only functions and 46% use it because the backend for a single-page software or devoted JavaScript frontend.
The steadiness is even clearer when builders are requested for his or her main method. Half primarily construct server-rendered functions, whereas 44% primarily use Django for both APIs or devoted JavaScript frontends.
The JavaScript numbers inform the sharper story. React has barely moved in 5 years — 37% in 2021, 38% in the present day. What’s modified is all the things round it: jQuery has fallen from 37% to 23%, Vue from 28% to 17%, whereas htmx has climbed from simply 5% to 34%. htmx isn’t taking share from React — it’s modernizing the server-rendered facet of the divide that jQuery used to personal.
That flexibility is certainly one of Django’s strengths. The identical framework can sit on the middle of a hypermedia software or behind an API consumed by React, cell apps, or different shoppers.
Django stays unusually comfy on each side of the frontend divide.
Conclusion
Throughout the survey, the sample is constant: builders are altering their instruments and workflows far quicker than they’re altering the framework beneath them. The elements of Django they worth most stay acquainted — fashions, the admin, authentication — and even deployment stays defiantly retro, with 54% transport monoliths and 44% self-hosting.
Even the place Django itself is evolving, it does so intentionally: 33% of respondents use its async options and one other 40% plan to — change supplied as an opt-in, not a rewrite.
That stability is more and more priceless. Builders can experiment with a brand new agent, exchange a number of instruments with Ruff or uv, add htmx to a template, or undertake a brand new kind checker with out having to rethink the framework beneath their software.
Django’s benefit in 2026 just isn’t that it’s the latest factor. It’s that it provides builders a strong, reliable — sure, boring — place from which to strive the most recent issues. Boring is so again.
PyCharm for Django Fundraiser
Get a brand new PyCharm Professional license or renew your present one at 30% off, with 100% of your buy quantity going to the Django Software program Basis.
Discover the whole 2026 Django Builders Survey Outcomes.
