n8n 3.0 is in beta, built for reliability over new features
The next major n8n release puts stability and security first, with a stable version due in early December.

A foundation release
n8n released the beta of version 3.0 on October 27, with the stable version expected in early December. Unlike most major releases, 3.0 is not about new features. It removes old features, tightens security defaults and reworks internals for reliability.
The reason is versioning. n8n only makes breaking changes in a major release, so a year of them has been bundled into 3.0. New features have kept arriving all along: n8n has shipped more than 45 releases since 2.0 in December 2025, roughly one a week.
Reliability changes
- In queue mode, manual executions always run on workers, taking that load off the main instance.
- Binary data can no longer be kept in memory; it goes to the filesystem or external storage such as S3.
- The Code node's default timeout drops from 5 minutes to 1 minute (still configurable).
- Limits on decompressing files are lower, to guard against oversized archives.
- The Cron and Interval nodes are replaced by the Schedule Trigger, which uses a more durable scheduler.
Security changes
- Encryption key rotation is always on.
- Slack credentials need a signing secret so incoming requests can be verified.
- The default list of blocked addresses for SSRF protection is longer.
- Unverified community packages are switched off by default.
Self-hosting is Docker-only
Self-hosted n8n now runs on Docker. Installing with npm or running npx n8n is no longer supported, partly because the new security features rely on containers. A one-line setup script replaces it for quick installs. Kubernetes and Helm setups are not affected.
What's removed
Several older nodes are gone: Function, Function Item and Item Lists, Cron and Interval, the legacy OpenAI node, older AI Agent modes, older versions of Edit Image, and nodes for services that no longer exist. Function is replaced by the Code node, and Item Lists by Split Out, Aggregate, Sort, Limit, Remove Duplicates or Summarize. Many workflows are migrated automatically.
Chat Hub is off by default but can be switched back on. Import from URL has been removed from the editor, and the Ask AI tab has been removed from the Code node.
How to upgrade
Version 2.39 and later includes a Migration Report under Settings. It lists every affected workflow and instance setting, fixes many issues automatically (saving each fix as a new version for you to review while the live workflow keeps running on 2.x), suggests an owner for each workflow and exports a CSV.
n8n's recommended path is: update to the latest 2.x, work through the Migration Report, take a full backup, then upgrade a test or staging instance before production. The upgrade is one-way; the only way back to 2.x is restoring a backup. Credentials and execution history carry over.
There is no direct jump from 1.x. Teams on 1.x must move to 2.x first, and support for 1.x ends on December 31, 2026. After 3.0 is stable, 2.x gets security and critical fixes for 12 months.
Why it matters for your business
n8n is the backbone of most automations we build, so this release affects our clients directly. Two dates matter: 1.x support ends on December 31, and 2.x moves to security fixes only once 3.0 is stable.
Most of the breaking changes are sensible, but they will catch out older workflows. Function nodes, Cron triggers and the legacy OpenAI node are common in workflows built a couple of years ago, and npm-based self-hosted installs will need moving to Docker.
We can audit your instance with the Migration Report, fix or retire what it flags, and run the upgrade on a copy first, so nothing breaks in production on switch-over day.
What to do next
- Check which n8n version you run; if it is 1.x, plan the move to 2.x before December 31
- Update to the latest 2.x and open Settings > Migration Report
- If you self-host with npm or npx, plan a move to Docker
- Take a full backup and test 3.0 on a copy before touching production

ConnectHub
