Nearly 1,800 developers from the Ruby on Rails community across 80 countries shared their thoughts on the tools, frameworks, and workflows that shape their daily development lives.
\nThis year, we also asked about something that's on everyone's mind: AI. How teams are using it, how they feel about it, and whether it's actually changing the way they work.
\nThis is the ninth edition of our survey since 2009. For nearly two decades, it has been one of the key ways our team at Planet Argon keeps a pulse on how this community has evolved.
\nSome of these questions date back to our very first survey, allowing us to track trends over time. Others are new additions, designed to reflect what's changed in our ecosystem since we last asked.
\nNow it’s time to dig in. We hope you find something useful, surprising, or at least validating here.
\nHow to read the charts: Responses were collected from April to early July 2026. Percentages are based on the people who answered each question. Many questions let people pick more than one answer, so those totals add up to more than 100%. Write-in answers were grouped by hand.
\nEach question lists its answers as Answer: count (percent). Percentages use the same base as the charts on the web page: the number of people who answered the question. For questions where people could pick more than one answer, that base is listed and the percentages add up to more than 100%. Trend questions are tables with one row per survey year. The link after each question opens it on the web page.
Who took the 2026 survey.
\nPercent of respondents.
\nBase: 1,783 respondents.
\n\nPercent of respondents by survey year.
\n| Survey year | CS degree | Bootcamp | Self-taught | On the job | Other |
|---|---|---|---|---|---|
| 2018 | 33% | 5% | 50% | 9% | 3% |
| 2020 | 36% | 6% | 45% | 9% | 4% |
| 2022 | 36% | 7% | 44% | 10% | 3% |
| 2024 | 36% | 7% | 44% | 10% | 3% |
| 2026 | 44.4% | 7.5% | 41.1% | 7% | – |
Values: percent of respondents.
\n\nPercent of respondents.
\nBase: 1,781 respondents.
\n\nCommunity insight from Amanda Perino
\n\nWhat stands out to me most is that only 7% of respondents are new to Rails. I don't necessarily read that as a lack of growth, but as a reminder that people who are newest to Rails may also be the least likely to know about a Rails community survey in the first place. Do they even realize they're part of the community yet?
\nIf an agent increasingly influences which framework someone first encounters, the challenge is no longer just helping people discover Rails, it's helping them discover the Rails community, too. How do we make sure those newcomers find the wealth of resources, support, history, and opportunities available to them?
\nThe doors to Rails are open wider than ever before, but we still need folks at the door welcoming people in. That's one of the biggest new opportunities agentic programming creates for the community.
\n— Amanda Perino, Executive Director, Rails Foundation
How Rails development teams are shaped and how they collaborate.
\nPercent of respondents.
\nBase: 1,781 respondents.
\n\nPercent of respondents by survey year.
\n| Survey year | Shrunk | About the same size | Grown |
|---|---|---|---|
| 2022 | 9% | 37% | 54% |
| 2024 | 18% | 46% | 36% |
| 2026 | 18.9% | 50.9% | 30.1% |
Values: percent of respondents.
\n\nPercent of respondents by survey year.
\n| Survey year | Mostly consumer-only (public) | A balanced mix of consumers and internal users | Mostly internal users only |
|---|---|---|---|
| 2009 | 30% | 58.7% | 11.3% |
| 2012 | 25.7% | 64.7% | 9.6% |
| 2014 | 24.1% | 66.9% | 9% |
| 2016 | 46.3% | 41.3% | 12.4% |
| 2018 | 42.3% | 46.1% | 11.6% |
| 2020 | 42.1% | 45% | 12.9% |
| 2022 | 43% | 47% | 10% |
| 2024 | 48% | 41% | 12% |
| 2026 | 47.2% | 42% | 10.8% |
Values: percent of respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,645 respondents. Counts are mentions; percentages are of respondents.
\n\nRespondents could pick up to three. Percent of respondents; totals add up to more than 100%.
\nBase: 1,437 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: Authentication and authorization setup
\n\nWhat versions the community is running and what holds upgrades back.
\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,796 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents by survey year. Select all that apply, so totals add up to more than 100%. mise appeared twice on the form; respondents who picked both are counted once.
\n| Survey year | RVM | rbenv | None... manual installs | chruby | asdf | mise | Docker (base images) | nix |
|---|---|---|---|---|---|---|---|---|
| 2016 | 40% | 34% | 22% | 4% | – | – | – | – |
| 2018 | 39% | 35% | 17% | 5% | – | – | – | – |
| 2020 | 35% | 43% | 3% | 7% | 9% | – | – | – |
| 2022 | 29% | 44% | 2% | 6% | 17% | – | – | – |
| 2024 | 17% | 33% | 1% | 6% | 21% | 2% | 18.9% | 0.8% |
| 2026 | 17.4% | 39.4% | 1.7% | 4.8% | 20.1% | 37.6% | 25.4% | 2.7% |
Values: percent of respondents.
\n\nCommunity insight from Jeff Dickey
\n\nMy career started with Ruby around 2010 and I remember using rvm and found the mechanics of how it worked fascinating. Ever since, I wanted to build a polyglot version manager that solved all of the problems I saw with dev environment setup. It's incredible to see my ideas having such an impact in the Ruby/Rails community. I also recently GA'ed precompiled Ruby binaries for mise and am happy to see that folks really seem to love them.
\n— Jeff Dickey, Creator of mise
Select all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,790 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents.
\nBase: 1,779 respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,236 respondents. Counts are mentions; percentages are of respondents.
\n\nCommunity insight from Ernesto Tagwerker
\n\nTwo years ago, in the Rails 2024 survey, before AI coding tools were part of daily work for most of us, the top reasons teams gave for not upgrading Rails or Ruby were basically the same as this year: "not enough time," "not enough budget," and "not a priority." That tells us something important. If AI tooling had ever been the missing piece, that list would have shifted by now. It hasn't, because this was never a tooling problem, it's an engineering culture problem.
\nAt FastRuby, we've seen firsthand AI noticeably speed up the mechanical parts of an upgrade: untangling old dependencies, addressing deprecation warnings, and triaging a flaky test suite. But the customers who come to us still running Rails 6 or Ruby 2.x weren't stuck because the tools didn't exist. They're carrying years of technical debt in dependencies nobody made time to address, and no amount of AI assistance fixes a roadmap that never made room for the upgrade. "Not enough time" and "no budget" are just more polite ways of saying "not a priority." Until that changes, neither will the survey results.
\n— Ernesto Tagwerker, Founder & CTO of FastRuby
How Rails applications are configured and what they integrate with.
\nPercent of respondents by survey year.
\n| Survey year | Monolith | Hybrid | Microservices |
|---|---|---|---|
| 2018 | 62% | 29% | 9% |
| 2020 | 68% | 25% | 8% |
| 2022 | 72% | 21% | 7% |
| 2024 | 77% | 17% | 6% |
| 2026 | 80.7% | 15.6% | 3.7% |
Values: percent of respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,363 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: MailPace (formerly OhMySMTP), SendPro (formerly Flowmailer)
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,666 respondents. Counts are mentions; percentages are of respondents.
\n\nCommunity insight from Rosa Gutiérrez
\n\nIt's been 2 years since Solid Queue 1.0 was launched and included by default in Rails. It's what you get when you run `rails new`. I'm very pleased to see that not everyone is automatically deleting it afterwards! 😅 Maybe there are even brave folks who are migrating their existing apps to it from another backend.
\nIn the last couple of years, all big features have been contributed to or heavily requested by the community: single process mode, dynamic recurring tasks, fibers, batches... we're close to version 2.0 now, and I hope it continues maturing and improving, and becoming a boring, reliable piece that people don't need to think about because it just works.
\n— Rosa Gutiérrez, Principal Programmer at 37signals
Percent of respondents.
\nBase: 1,742 respondents.
\n\nAsked of respondents with a companion native app. Select all that apply. Percent of those respondents; totals add up to more than 100%.
\nBase: 635 respondents. Counts are mentions; percentages are of respondents.
\n\nHow Rails developers built their skills and stay connected to the community.
\nPercent of respondents.
\nBase: 1,787 respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,785 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents.
\nBase: 1,666 respondents.
\n\nCommunity insight from Amanda Perino
\n\nWhat a difference two years makes. The last community survey feels like another lifetime! Almost overnight, agents have become one of the first resources developers reach for when researching, learning, and building, so they've become an important audience for the Rails Foundation, too. We now need to make discovering, learning, and embracing Rails as seamless for agents as it is for humans.
\nFor the Rails Foundation, the mission remains the same- we still want to make sure someone learns about Rails (wherever they ask the question) and gets quickly from idea to building a working MVP. But it requires a shift in strategy and how we think about documentation, education, and marketing for both humans and agents.
\nWe have a lot of work to do in this new agentic world, but the opportunity for Rails is huge, so we're excited about what the future holds.
\n— Amanda Perino, Executive Director, Rails Foundation
Write-in responses, listed by number of mentions. Percentages are of respondents who answered this question.
\nMost Mentioned
\nOther Mentions
\nRespondents could pick up to three. Percent of respondents; totals add up to more than 100%.
\nBase: 1,770 respondents. Counts are mentions; percentages are of respondents.
\n\nCommunity insight from John Athayde
\n\nThe Rails Guides were essential in my growth in the framework over the past two decades and I loved being able to work with the docs team and foundation to reimagine and update them for a new generation of Rails programmers. The knowledge isn't just about how to do something but why we do things in Rails and the convention that it provides giving us (and agents!) the power to write and ship effective software that's still easy to read and understand.
\n— John Athayde, Owner, Meticulous
Percent of respondents by survey year.
\n| Survey year | Yes | No |
|---|---|---|
| 2016 | 30% | 70% |
| 2018 | 24% | 76% |
| 2020 | 23% | 77% |
| 2022 | 23% | 77% |
| 2024 | 26% | 74% |
| 2026 | 28% | 72% |
Values: percent of respondents.
\n\nWrite-in responses, ranked by number of mentions. Rails World led by a wide margin: 30% of respondents plan to attend.
\nTop 10
\nOther Mentions
\nPercent of respondents who said yes, by survey year. Local in-person and online groups were asked as separate questions. Online groups were first asked in 2020.
\n| Survey year | Online | Local in person |
|---|---|---|
| 2016 | – | 45.6% |
| 2018 | – | 37.9% |
| 2020 | 47% | 34.8% |
| 2022 | 44.3% | 22.7% |
| 2024 | 40.5% | 27.4% |
| 2026 | 34.8% | 26.3% |
Values: percent of respondents.
\n\nCommunity insight from Débora Fernandes
\n\nIt's amazing to see the Ruby/Rails community more connected and active than ever in recent years! Unlike the last decade, when events were mostly local, it feels like borders have simply disappeared. We've truly become a global community of Ruby Friends!
\nAs an attendee, it's exciting to arrive at an event knowing that, beyond the talks, the conversations I'll have in the hallways will be just as enriching, thanks to the diversity of people and perspectives we have today.
\nAnd especially as an organizer of larger events like Tropical on Rails 24, and local and consistent ones like Rails Girls São Paulo a project that brings and encourages women to code, seeing Brazil grow in the community's top of mind is incredibly exciting. The numbers around people's intention to visit Brazil make this even more special. I also think this representation comes from all the people who have been organizing and nurturing local events here since Rails was launched. That work over the years has helped build a very strong and welcoming community in Brazil.
\nThe Brazilian vibe, at home or abroad, has been winning people over in recent years, and I believe we'll see more and more people visiting Brazil, joining our events, and speaking at them!
\nI hope to see our welcome DNA become an increasingly strong part of our wonderfully diverse community!
\n— Débora Fernandes, Débora is a software engineer and has been working with Rails since 2012. She is one of the Rails Girls São Paulo organizers, and has spent years helping grow the local Ruby community through events, mentoring, and teaching programming.
Top 10
\nHonorable Mentions
\n\n\nCommunity insight from Robby Russell
\n\nHonestly, I asked our team on multiple occasions to double- and triple-check this data. I knew this was a write-in question, so seeing both Maintainable and On Rails show up so prominently seemed a little suspicious.
\nTurns out… the data is accurate.
\nIt’s an honor to host two podcasts with their own particular flavors. Maintainable lets me talk with people about the messy, long-term realities of building and maintaining software. On Rails is where guests share how Ruby on Rails is part of the secret sauce behind real-world applications. At the time of this survey, On Rails had only been on the air for a little over a year.
\nIt’s a delight to know these conversations are connecting with people, and a privilege to see both shows mentioned alongside so many other talented podcasts in the Ruby and Rails community.
\nThanks to everyone who listens, to our real and occasionally not-so-real sponsors, and most importantly, to the guests who keep showing up to talk shop with us.
\n— Robby Russell, Co-founder and CEO of Planet Argon, a Portland consultancy that has been helping teams maintain and upgrade Rails apps they didn't write since 2005. He created Oh My Zsh and hosts the Maintainable and On Rails podcasts.
Top Ten
\nHonorable Mentions
\n\n\nCommunity insight from Peter Cooper
\n\nIt's said that eighty percent of success is showing up and apparently showing up (almost) every week for 16 years has sweet rewards, like being at the top of this chart! I want to dedicate this achievement to the Rubyists who are no longer with us but were friends or influences along the way, including Jason Seifer, Jim Weirich, and Noah Gibbs.
\n— Peter Cooper, Peter Cooper is the founder of Cooperpress, the team behind Ruby Weekly and JavaScript Weekly.
Community insight from Irina Nazarova
\n\nEvil Martians blog is the heart of our team. We write not only to share learnings with the big tech community, but to share them with each other too! This process is contagious, and it works thanks to all the support we are getting from our Editor-in-Chief, Travis Turner.
\n— Irina Nazarova, CEO, Evil Martians
Top 10
\nHonorable Mentions
\n\n\nTop 12
\nCommunity contribution and the gem ecosystem.
\nPercent of respondents.
\nBase: 1,776 respondents.
\n\nWrite-in responses, ranked by number of mentions. Percent of respondents who answered this question. Tied answers share a rank.
\nTop 10
\nOther Responses
\nWrite-in responses, ranked by number of mentions. Percent of respondents who answered this question. Tied answers share a rank.
\nTop 10
\nOther Responses
\nCommunity insight from Carmine Paolino
\n\nRuby was the first programming language I fell in love with. Its relentless focus on beauty, aesthetics, taste, and the joy of programming made me feel seen. People cared about how code felt to read and write. I missed that.
\nWhen I returned at the end of 2024 to build Chat with Work, I couldn't believe how familiar Rails still felt. Fifteen years later, Active Record still looked like Active Record. I could come back and build things. I wanted that same feeling from my AI code, and RubyLLM grew out of that desire.
\nFinding RubyLLM among the five most-loved gems is an incredible welcome home. What started as something I needed for my own application has grown through people building things I hadn't imagined, sharing ideas, and contributing. To everyone who wrote it in: thank you. I hope RubyLLM gives you some of the same enjoyment that brought me back to Ruby.
\n— Carmine Paolino, Creator of RubyLLM, founder of Chat with Work, and Berlin.rb organizer
How JavaScript sits alongside Rails in modern applications.
\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,692 respondents. Counts are mentions; percentages are of respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,705 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: Preact 1%, Elm 1%, Unpoly 0.4%
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,284 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: WebdriverIO, Ava
\n\nTesting, monitoring, and staying on top of vulnerabilities.
\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,777 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents.
\nBase: 1,745 respondents.
\n\nPercent of respondents.
\nBase: 1,718 respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,568 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: Sumo
\n\nCommunity insight from Nate Berkopec
\n\nIt's not hard to see why Sentry has become so popular. It's a great performance monitor, and you also get great error monitoring, frontend RUM, monitors, metrics and more all in the same place. In the past 2 years it's become a full o11y platform, and it's easy to use for small teams. For larger enterprise orgs, Datadog is still great. Also, it's not a coincidence: the top two options also have the best MCPs of all their competition.
\n— Nate Berkopec, Nate Berkopec is the owner of Speedshop, a Ruby on Rails performance consultancy. He is the author of several books on Ruby and Rails performance.
Select all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,720 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: Coveralls, Skunk
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,680 respondents. Counts are mentions; percentages are of respondents.
\n\nRespondents could pick up to three. Percent of respondents; totals add up to more than 100%.
\nBase: 1,671 respondents. Counts are mentions; percentages are of respondents.
\n\nWhere Rails apps run and how they get there.
\nPercent of respondents by survey year.
\n| Survey year | Multiple times a day | Multiple times a week | Multiple times a month | Multiple times a quarter |
|---|---|---|---|---|
| 2009 | 13% | 29% | 38% | – |
| 2012 | 26% | 33% | 29% | – |
| 2014 | 27% | 35% | 26% | – |
| 2016 | 26% | 34% | 32% | – |
| 2018 | 34% | 35% | 26% | – |
| 2020 | 36.9% | 33.2% | 24.2% | – |
| 2022 | 39% | 31% | 24% | – |
| 2024 | 37% | 36% | 20% | 5% |
| 2026 | 41.2% | 36.2% | 17.9% | 4.7% |
Values: percent of respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,773 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents by survey year. Select all that apply, so totals add up to more than 100%. Kamal grew from 2.4% in 2024 to 23.8% in 2026, after Rails 8 made it the default deployment tool for new apps.
\n| Survey year | Continuous Deployment via CI | Git | Kamal | Capistrano | None |
|---|---|---|---|---|---|
| 2016 | 28.4% | 52.4% | – | 64.6% | 6.8% |
| 2018 | 35.1% | 59.4% | – | 49.1% | 7.5% |
| 2020 | 51.9% | 37.1% | – | 34.4% | 8.4% |
| 2022 | 57.2% | 37.5% | – | 25.3% | 7.6% |
| 2024 | 57.4% | 33.3% | 2.4% | 17.5% | 8.6% |
| 2026 | 59.9% | 39.9% | 23.8% | 15.5% | 7.5% |
Values: percent of respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,669 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: Hatchbox, Railway, Coolify, Vultr
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,718 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents by survey year. Select all that apply, so totals add up to more than 100%. 4% use no CI.
\n| Survey year | Github Actions | Circle CI | Gitlab | Jenkins | Buildkite | Semaphore | Heroku CI |
|---|---|---|---|---|---|---|---|
| 2016 | – | 26% | – | 24% | – | 8% | – |
| 2018 | – | 30.7% | 14.9% | 23.9% | – | 7.6% | – |
| 2020 | – | 30.6% | 18.7% | 16.9% | – | 6.9% | – |
| 2022 | 43.2% | 28.6% | 16.7% | 13% | – | 5% | – |
| 2024 | 54.8% | 19.6% | 14.4% | 7.1% | 4.7% | 3.1% | 4.8% |
| 2026 | 69.3% | 16.2% | 14.5% | 6.8% | 5% | 3.5% | 3% |
Values: percent of respondents.
\nRunner-ups in 2026: CodePipeline (AWS) 2%, Bitbucket Pipelines 2%, Travis CI 0.4%, Codeship 0.1%
\n\nPercent of respondents by survey year. Select all that apply, so totals add up to more than 100%. 22% don't use a CDN.
\n| Survey year | CloudFlare | CloudFront | Cloudinary | Google Cloud CDN | Fastly | BunnyCDN | Akamai |
|---|---|---|---|---|---|---|---|
| 2018 | 30% | 57% | – | – | – | – | 5% |
| 2020 | 35% | 51% | – | – | 7% | – | 2% |
| 2022 | 41% | 46% | – | – | 7% | – | 2% |
| 2024 | 35% | 28% | 3% | 4% | 3% | 1% | 1% |
| 2026 | 45.4% | 32.2% | 3.6% | 3.3% | 2.5% | 2% | 1.1% |
Values: percent of respondents.
\nRunner-ups in 2026: Azure CDN 1%
\n\nWhich data stores back Rails applications.
\nPercent of respondents by survey year. Select all that apply, so totals add up to more than 100%.
\n| Survey year | Redis | PostgreSQL (via JSONB / HStore) | ElasticSearch | MongoDB | DynamoDB | Neo4j | Cassandra |
|---|---|---|---|---|---|---|---|
| 2016 | 69.8% | – | 30.8% | 30.1% | – | 2.4% | 3.5% |
| 2018 | 71.6% | – | 33.3% | 25.4% | – | 2.6% | 4.1% |
| 2020 | 64.7% | – | 29.3% | 15.1% | – | 1.7% | 2.3% |
| 2022 | 64.4% | – | 26.9% | 15% | – | 1.2% | 1.4% |
| 2024 | 63.6% | 27.1% | 20.8% | 7.5% | 5.3% | 1% | 0.9% |
| 2026 | 76.3% | 45.9% | 25.4% | 8.8% | 7.6% | 1.1% | 0.5% |
Values: percent of respondents.
\nRunner-ups in 2026: CouchDB 0.4%
\nNo longer asked: Riak
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,774 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents by survey year.
\n| Survey year | PostgreSQL | SQLite | MySQL | MariaDB |
|---|---|---|---|---|
| 2016 | 83.9% | – | 14.1% | – |
| 2018 | 86.9% | – | 11.8% | – |
| 2020 | 84.6% | – | 8.9% | – |
| 2022 | 85% | – | 8.7% | – |
| 2024 | 81.9% | 3.9% | 7.3% | 1.3% |
| 2026 | 84.5% | 7.3% | 5.8% | 1.9% |
Values: percent of respondents.
\n\nThe tools developers reach for on their local machines.
\nPercent of respondents.
\nBase: 1,756 respondents.
\n\nPercent of respondents.
\nBase: 1,744 respondents.
\n\nPercent of respondents by survey year. Zed (10.2%) and Cursor (5.4%) were write-in responses in 2026 — they weren't offered as formal options in prior surveys.
\n| Survey year | Visual Studio Code | Vim-based | RubyMine | Zed | Sublime Text | Cursor | Emacs |
|---|---|---|---|---|---|---|---|
| 2016 | – | 28.1% | – | – | 37.7% | – | 3.2% |
| 2018 | – | 30.4% | – | – | 35.4% | – | 5.3% |
| 2020 | 31.2% | 20.2% | 15% | – | 16.1% | – | 3.4% |
| 2022 | 42% | 17% | 18.2% | – | 11.1% | – | 2.7% |
| 2024 | 43.4% | 19.2% | 18.4% | – | 7% | – | 3.3% |
| 2026 | 44.7% | 23% | 18.9% | 10.2% | 6.8% | 5.4% | 3.4% |
Values: percent of respondents.
\nRunner-ups in 2026: TextMate 1%, IntelliJ 1%
\nNo longer asked: Atom
\n\nCommunity insight from Robby Russell
\n\nI have a feeling this answer will look very different the next time we run this survey. The line between code editor, IDE, and AI coding environment is getting increasingly blurry. Maybe by then we’ll all be vibe coding our own editors. 🤷
\n— Robby Russell, Co-founder and CEO of Planet Argon, a Portland consultancy that has been helping teams maintain and upgrade Rails apps they didn't write since 2005. He created Oh My Zsh and hosts the Maintainable and On Rails podcasts.
How AI tools are (and aren't) landing in Rails workflows.
\nPercent of respondents.
\nBase: 1,764 respondents.
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,716 respondents. Counts are mentions; percentages are of respondents.
\nOther responses: Windsurf, Amazon Q Developer, Cline, Aider, Tabnine, Continue.dev
\n\nSelect all that apply. Percent of respondents; totals add up to more than 100%.
\nBase: 1,692 respondents. Counts are mentions; percentages are of respondents.
\n\nPercent of respondents.
\nBase: 1,756 respondents.
\n\nPercent of respondents.
\nBase: 1,753 respondents.
\n\nPercent of respondents.
\nBase: 1,746 respondents.
\n\nPercent of respondents.
\nBase: 1,754 respondents.
\n\nPercent of respondents.
\nBase: 1,755 respondents.
\n\nPercent of respondents.
\nBase: 1,752 respondents.
\n\nPercent of respondents.
\nBase: 1,722 respondents.
\n\nPercent of respondents.
\nBase: 1,720 respondents.
\n\nPercent of respondents.
\nBase: 1,708 respondents.
\n\nPercent of respondents.
\nBase: 1,708 respondents.
\n\nPercent of respondents.
\nBase: 1,714 respondents.
\n\nPercent of respondents.
\nBase: 1,707 respondents.
\n\nWrite-in responses, grouped by hand. Percent of respondents who answered; people could name more than one.
\nBase: 1,146 respondents. Counts are mentions; percentages are of respondents.
\nOther Responses: Claude (unspecified), GLM / Big Pickle (Zhipu AI), Claude Haiku, MiniMax, Gemma (Google), GitHub Copilot, Grok (xAI), Mistral/Devstral, Local LLMs, Amp
\n\nSentiment on where Rails is heading and what teams are planning next.
\nPercent of respondents.
\nBase: 1,736 respondents.
\n\nTwo questions: what significant technical changes has your team completed in the last 24 months, and what is it planning or evaluating in the next ~12 months? These reflect choices individual teams made for their own codebases and constraints, not recommendations. Select all that apply. Percent of respondents who answered each question (1,659 and 1,509). None in the last 24 months: 5%. None planned: 11%.
\n| Change | Completed, last 24 months | Planning or evaluating, next 12 months |
|---|---|---|
| Upgrades: Rails version upgrade | 75.5% | 62% |
| Upgrades: Ruby version upgrade | 70.8% | 57.6% |
| AI & the app layer: Adopted AI agents or agentic coding workflows | 40.9% | 19% |
| AI & the app layer: Overhauled the JavaScript / frontend layer | 20.4% | 18.2% |
| AI & the app layer: Replaced or overhauled background job processing | 15.3% | 9.6% |
| AI & the app layer: Replaced or rewrote the authentication system | 10% | 8.2% |
| Infrastructure & deploys: Adopted Kamal | 13.5% | 7.3% |
| Infrastructure & deploys: Adopted self-hosted CI runners | 7.6% | 4.4% |
| Infrastructure & deploys: Consolidated CI pipelines around GitHub Actions | 14.5% | 4.2% |
| Infrastructure & deploys: Migrated away from Heroku | 10.8% | 12.5% |
| Infrastructure & deploys: Moved to self-hosted infrastructure | 7.3% | 4.4% |
| Infrastructure & deploys: Switched to a different hosting provider | 13.7% | 8.5% |
| Architecture: Consolidated services back into a monolith | 5.7% | 3.6% |
| Architecture: Extracted functionality into Rails Engines | 5.8% | 4.8% |
| Architecture: Introduced Packwerk for modular boundaries | 3.3% | 2.5% |
| Architecture: Merged multiple apps into one | 6.9% | 4.6% |
| Architecture: Migrated to a monorepo | 4.5% | 1.7% |
| Architecture: Moved from monolith toward services | 4.8% | 5.8% |
| Data: Migrated away from SQLite | 1.1% | 0.5% |
| Data: Migrated to a different database engine | 5.1% | 2.5% |
| Data: Migrated to SQLite in production | 2.8% | 1.5% |
| Testing & types: Adopted RBS / Steep for type checking | 1.3% | 1.8% |
| Testing & types: Adopted Sorbet for static type checking | 2.1% | 1.6% |
| Testing & types: Migrated from Minitest to RSpec | 1.9% | 1.1% |
| Testing & types: Migrated from RSpec to Minitest | 5.5% | 4% |
| Testing & types: Shifted to running tests locally before deploying | 4.2% | 3.8% |
Percent of respondents.
\nBase: 1,713 respondents.
\n\nPercent of respondents.
\nBase: 1,703 respondents.
\n\nPercent of respondents.
\nBase: 1,708 respondents.
\n\nAI is no longer a sidebar in Rails development. It showed up throughout this year's responses: in the tools people reach for, the gems they love, and the workflows they're rethinking. The community isn't watching from a distance; we're actively exploring.
\nSome things remain constant. A few of the same tools, books, and podcasts that have anchored this ecosystem for years are still showing up. And new ones are appearing alongside them.
\nAt Planet Argon, we remain steadfast in our belief that Ruby on Rails is an exceptional platform to invest in, and we're excited to continue supporting its growth. If a few of these numbers hit close to home (inherited code, an upgrade that keeps sliding, an AI policy that's still a draft), that's the work we do every day. Come say hello.
\nThank you to everyone who took the time to contribute to the 2026 Rails Community Survey, our ninth edition since 2009. Eighty countries' worth of perspectives on a framework that keeps finding new reasons to stick around. We'll see you in 2028!
","headings":[{"level":2,"text":"The Results Are In!","id":"the-results-are-in"},{"level":2,"text":"How to read this document","id":"how-to-read-this-document"},{"level":2,"text":"Demographics","id":"demographics"},{"level":3,"text":"Which one of the following applies to you?","id":"which-one-of-the-following-applies-to-you"},{"level":3,"text":"How did you learn to code?","id":"how-did-you-learn-to-code"},{"level":3,"text":"How many years of experience do you have developing with Ruby on Rails?","id":"how-many-years-of-experience-do-you-have-developing-with-ruby-on"},{"level":2,"text":"Teams & Responsibilities","id":"teams-responsibilities"},{"level":3,"text":"How big is the software development team for your primary Rails app(s)?","id":"how-big-is-the-software-development-team-for-your-primary-rails-"},{"level":3,"text":"Has your development team grown or shrunk in the last two years?","id":"has-your-development-team-grown-or-shrunk-in-the-last-two-years"},{"level":3,"text":"Who uses your application/s?","id":"who-uses-your-application-s"},{"level":3,"text":"How do you give your developers access to realistic data for local development and bug reproduction?","id":"how-do-you-give-your-developers-access-to-realistic-data-for-loc"},{"level":3,"text":"When new developers join your project, where do they tend to struggle most?","id":"when-new-developers-join-your-project-where-do-they-tend-to-stru"},{"level":2,"text":"Ruby & Rails Version Updates","id":"ruby-rails-version-updates"},{"level":3,"text":"What versions of Ruby are you using in your applications?","id":"what-versions-of-ruby-are-you-using-in-your-applications"},{"level":3,"text":"Which Ruby version manager are you using?","id":"which-ruby-version-manager-are-you-using"},{"level":3,"text":"What versions of Rails are you using in your applications?","id":"what-versions-of-rails-are-you-using-in-your-applications"},{"level":3,"text":"How long has your Ruby on Rails application been in development?","id":"how-long-has-your-ruby-on-rails-application-been-in-development"},{"level":3,"text":"What keeps you from keeping your Ruby and Rails versions up-to-date?","id":"what-keeps-you-from-keeping-your-ruby-and-rails-versions-up-to-d"},{"level":2,"text":"Configuration","id":"configuration"},{"level":3,"text":"What are you primarily working on?","id":"what-are-you-primarily-working-on"},{"level":3,"text":"What third-party email services do your Rails apps integrate with?","id":"what-third-party-email-services-do-your-rails-apps-integrate-wit"},{"level":3,"text":"Which Active Job adapters do you use to process background jobs?","id":"which-active-job-adapters-do-you-use-to-process-background-jobs"},{"level":3,"text":"Do any of your Rails applications have a companion native mobile app?","id":"do-any-of-your-rails-applications-have-a-companion-native-mobile"},{"level":3,"text":"How is your companion mobile app built?","id":"how-is-your-companion-mobile-app-built"},{"level":2,"text":"Learning","id":"learning"},{"level":3,"text":"What was your level of web programming expertise before learning Ruby on Rails?","id":"what-was-your-level-of-web-programming-expertise-before-learning"},{"level":3,"text":"In the first two years of learning Ruby on Rails, what methods did you use?","id":"in-the-first-two-years-of-learning-ruby-on-rails-what-methods-di"},{"level":3,"text":"Of these methods of learning, which did you feel worked best for you?","id":"of-these-methods-of-learning-which-did-you-feel-worked-best-for-"},{"level":3,"text":"If you attended a bootcamp and/or school for learning to code...which one(s)?","id":"if-you-attended-a-bootcamp-and-or-school-for-learning-to-code-wh"},{"level":3,"text":"What type(s) of educational content helps you learn the most effectively?","id":"what-type-s-of-educational-content-helps-you-learn-the-most-effe"},{"level":3,"text":"Have you been to a conference (regional or bigger) for Ruby and/or Rails in the last year?","id":"have-you-been-to-a-conference-regional-or-bigger-for-ruby-and-or"},{"level":3,"text":"What conference(s) are you hoping to attend in the coming year?","id":"what-conference-s-are-you-hoping-to-attend-in-the-coming-year"},{"level":3,"text":"Do you participate in Ruby/Rails user groups?","id":"do-you-participate-in-ruby-rails-user-groups"},{"level":3,"text":"What are your favorite Rails-related technical podcasts?","id":"what-are-your-favorite-rails-related-technical-podcasts"},{"level":3,"text":"What are your favorite Rails-related newsletters and blogs?","id":"what-are-your-favorite-rails-related-newsletters-and-blogs"},{"level":3,"text":"Who are your favorite live coding streamers or screencasts?","id":"who-are-your-favorite-live-coding-streamers-or-screencasts"},{"level":3,"text":"What Rails-related books do you recommend others read?","id":"what-rails-related-books-do-you-recommend-others-read"},{"level":2,"text":"Gems & Open Source","id":"gems-open-source"},{"level":3,"text":"How often do you contribute to open source projects?","id":"how-often-do-you-contribute-to-open-source-projects"},{"level":3,"text":"Which Ruby gems frustrate you the most?","id":"which-ruby-gems-frustrate-you-the-most"},{"level":3,"text":"Which Ruby gems do you love?","id":"which-ruby-gems-do-you-love"},{"level":2,"text":"JavaScript + Rails","id":"javascript-rails"},{"level":3,"text":"What tool(s) do you use to manage your JavaScript libraries?","id":"what-tool-s-do-you-use-to-manage-your-javascript-libraries"},{"level":3,"text":"Which JavaScript libraries or frameworks do you use with Rails? Check all that apply.","id":"which-javascript-libraries-or-frameworks-do-you-use-with-rails-c"},{"level":3,"text":"What JavaScript testing frameworks are you using to write tests?","id":"what-javascript-testing-frameworks-are-you-using-to-write-tests"},{"level":2,"text":"Code Integrity & Security","id":"code-integrity-security"},{"level":3,"text":"What testing frameworks are you using to test your Rails app(s)?","id":"what-testing-frameworks-are-you-using-to-test-your-rails-app-s"},{"level":3,"text":"What is your current code to test/spec ratio?","id":"what-is-your-current-code-to-test-spec-ratio"},{"level":3,"text":"What is your go-to Ruby debugger tool?","id":"what-is-your-go-to-ruby-debugger-tool"},{"level":3,"text":"Which application monitoring services do you use?","id":"which-application-monitoring-services-do-you-use"},{"level":3,"text":"What code quality tools do you use?","id":"what-code-quality-tools-do-you-use"},{"level":3,"text":"What tools are you using to track security vulnerabilities?","id":"what-tools-are-you-using-to-track-security-vulnerabilities"},{"level":3,"text":"Where does your primary Rails application feel most fragile right now?","id":"where-does-your-primary-rails-application-feel-most-fragile-righ"},{"level":2,"text":"Deployment & DevOps","id":"deployment-devops"},{"level":3,"text":"How often do you deploy your Rails applications to production?","id":"how-often-do-you-deploy-your-rails-applications-to-production"},{"level":3,"text":"Where is your source code hosted?","id":"where-is-your-source-code-hosted"},{"level":3,"text":"Which automated deployment tools do you use? (select all that apply)","id":"which-automated-deployment-tools-do-you-use-select-all-that-appl"},{"level":3,"text":"Where do you host your production Rails app(s)?","id":"where-do-you-host-your-production-rails-app-s"},{"level":3,"text":"Which Rails Servers are you using in production?","id":"which-rails-servers-are-you-using-in-production"},{"level":3,"text":"Which Continuous Integration (CI) servers do you use? (select all that apply)","id":"which-continuous-integration-ci-servers-do-you-use-select-all-th"},{"level":3,"text":"What CDN(s) do you use? (select all that apply)","id":"what-cdn-s-do-you-use-select-all-that-apply"},{"level":2,"text":"Databases","id":"databases"},{"level":3,"text":"Which nosql databases do you use in production? (select all that apply)","id":"which-nosql-databases-do-you-use-in-production-select-all-that-a"},{"level":3,"text":"Which relational databases do you use in production?","id":"which-relational-databases-do-you-use-in-production"},{"level":3,"text":"Which database would you prefer to use in production?","id":"which-database-would-you-prefer-to-use-in-production"},{"level":2,"text":"OS, Editors & Servers","id":"os-editors-servers"},{"level":3,"text":"Which operating system do you primarily use for local development?","id":"which-operating-system-do-you-primarily-use-for-local-developmen"},{"level":3,"text":"Are you using Docker for local development?","id":"are-you-using-docker-for-local-development"},{"level":3,"text":"What is your preferred code editor?","id":"what-is-your-preferred-code-editor"},{"level":2,"text":"AI Adoption","id":"ai-adoption"},{"level":3,"text":"How is your team currently using AI tools in your development workflow?","id":"how-is-your-team-currently-using-ai-tools-in-your-development-wo"},{"level":3,"text":"Which AI coding tools or assistants do you use?","id":"which-ai-coding-tools-or-assistants-do-you-use"},{"level":3,"text":"Beyond coding assistance, how is your team exploring AI in relation to your Rails applications?","id":"beyond-coding-assistance-how-is-your-team-exploring-ai-in-relati"},{"level":3,"text":"AI coding tools are overhyped and won't fundamentally change how I build software.","id":"ai-coding-tools-are-overhyped-and-won-t-fundamentally-change-how"},{"level":3,"text":"I feel frustrated by the growing expectation to use AI in my workflow.","id":"i-feel-frustrated-by-the-growing-expectation-to-use-ai-in-my-wor"},{"level":3,"text":"I intentionally limit AI to boilerplate tasks so I stay in control of the real problem-solving.","id":"i-intentionally-limit-ai-to-boilerplate-tasks-so-i-stay-in-contr"},{"level":3,"text":"The pace of AI advancement makes me anxious about the long-term value of my skills.","id":"the-pace-of-ai-advancement-makes-me-anxious-about-the-long-term-"},{"level":3,"text":"AI tools are here to stay and I've made peace with that, limitations and all.","id":"ai-tools-are-here-to-stay-and-i-ve-made-peace-with-that-limitati"},{"level":3,"text":"I've adapted how I work and AI genuinely makes me more capable.","id":"i-ve-adapted-how-i-work-and-ai-genuinely-makes-me-more-capable"},{"level":3,"text":"Leadership largely ignores AI's impact on how we build software.","id":"leadership-largely-ignores-ai-s-impact-on-how-we-build-software"},{"level":3,"text":"Our team restricts or prohibits AI tools due to risk concerns.","id":"our-team-restricts-or-prohibits-ai-tools-due-to-risk-concerns"},{"level":3,"text":"AI use is limited to experiments; it hasn't made it into our real workflows.","id":"ai-use-is-limited-to-experiments-it-hasn-t-made-it-into-our-real"},{"level":3,"text":"Our approach to AI adoption feels reactive and poorly thought through.","id":"our-approach-to-ai-adoption-feels-reactive-and-poorly-thought-th"},{"level":3,"text":"We're taking AI seriously and working toward a realistic, balanced stance.","id":"we-re-taking-ai-seriously-and-working-toward-a-realistic-balance"},{"level":3,"text":"We have clear policies that actually empower engineers to use AI effectively.","id":"we-have-clear-policies-that-actually-empower-engineers-to-use-ai"},{"level":3,"text":"Which AI models do you reach for most in your development work?","id":"which-ai-models-do-you-reach-for-most-in-your-development-work"},{"level":2,"text":"The Future of Rails","id":"the-future-of-rails"},{"level":3,"text":"I feel confident security vulnerabilities are being addressed in new Rails releases.","id":"i-feel-confident-security-vulnerabilities-are-being-addressed-in"},{"level":3,"text":"What technical changes have teams taken on, or are considering?","id":"what-technical-changes-have-teams-taken-on-or-are-considering"},{"level":3,"text":"Ruby on Rails is my server side framework of choice.","id":"ruby-on-rails-is-my-server-side-framework-of-choice"},{"level":3,"text":"The Rails core team is shepherding the project in the right direction.","id":"the-rails-core-team-is-shepherding-the-project-in-the-right-dire"},{"level":3,"text":"I recommend that new developers learn to build Rails apps in 2026.","id":"i-recommend-that-new-developers-learn-to-build-rails-apps-in-202"},{"level":2,"text":"In Conclusion","id":"in-conclusion"}]}}