Finally, Real-Time Collaboration will arrive in WordPress 7.1 and not 7.0, given the performance issues and the changes needed for its new implementation.
Remember that you can listen to this program from Pocket Casts, Spotify, and Apple Podcasts or subscribe to the feed directly.
Program transcript
Hello, I’m Alicia Ireland, and you’re listening to WPpodcast, bringing the weekly news from the WordPress Community.
In this episode, you’ll find the information from May 4 to 10, 2026.
The big news this week around WordPress 7.0 is that real-time collaboration will not be in the final release. Matt Mullenweg made the decision to remove it, citing attack surface problems, race conditions, server load, memory efficiency, and recurring bugs found in fuzz testing. It’s a difficult call given the work invested, but it’s made in the interest of shipping a stable release. The feature will be re-evaluated during the WordPress 7.1 cycle, and in the meantime remains available through the Gutenberg plugin for anyone who wants to keep testing it.
The technical data behind that decision has also been published. Eight different hosting environments, including configurations with and without persistent object caching, participated in the performance tests. The analysis makes it fairly clear that a dedicated table storage strategy combined with object caching for user presence detection is the winner: roughly 52% faster than the current implementation and the only approach that scales well both with and without cache. This information will guide work on the next iteration of the feature.
With all of that in place, on May 8 WordPress 7.0 Release Candidate 3 was published, which already incorporates the removal of real-time collaboration and resolves over 143 additional issues since RC2. The final release date remains May 20. If you have plugins or themes, now is the time to test them against this RC3 and update the “Tested up to” field to 7.0.
Gutenberg 23.1 has arrived with a solid amount of updates, though many are still experimental. The most noticeable for everyday use is the improvement in image uploads: thumbnails are now generated in parallel rather than sequentially, which is especially noticeable when batch-uploading images through the gallery block or on slow connections.
In the experimental department there are interesting things. Managing custom taxonomies from the admin without writing PHP is one of the most anticipated: with the experiment enabled, a Taxonomies screen appears in Settings where you can create, edit, and delete taxonomies from the interface.
There is also a new media editor with a free-form image cropper for the image and site logo blocks. And the experiment to disable TinyMCE has evolved: instead of removing it entirely, it now simply hides the Classic block, leaving existing instances to keep working — which is far more reasonable.
The AI plugin continues its two-week release cadence: version 0.9.0 just shipped with content resizing, comment moderation, and various fixes, with the goal of hitting version 1.0.0 to coincide with WordPress 7.0. In parallel, the MCP adapter — which lets assistants like Claude or ChatGPT interact with WordPress — is finishing the last details before being published as a standalone plugin in the repository.
One of the most interesting debates centers on API key security. The team is evaluating a connector approval mechanism that would let administrators control which plugins can access connections configured with providers like OpenAI, Anthropic, or Google. Several contributors warn that solving this only from a plugin can give a false sense of security, and that what WordPress really needs at its core is a secrets management API in core. The idea is for this experiment to serve as a first step and push that conversation toward WordPress 7.1.
The team has also started defining what comes next after completing the technical foundations of AI in WordPress. The conversations point toward education, best practices, supporting other contributor teams with AI tools, and long-term strategic positioning.
The Developer Blog has published a practical guide to getting started writing end-to-end tests for WordPress using Playwright. The article starts from the premise that unit tests cover one layer, but E2E tests cover a different one: they simulate what a real user would do in the browser, checking how the whole system works, not isolated components. Gutenberg has been using them from the beginning, and for a few years they have also been part of WordPress Core.
The bulk of the article is oriented toward plugin and theme developers, with concrete examples from a book review project. For anyone who has never touched Playwright in the WordPress context, the guide is a good entry point: the ramp-up curve is reasonable, the examples are real, and the complete code is available in a public repository.
The Accessibility team has updated the requirements for obtaining the “accessibility-ready” label on themes in the official directory. The previous criteria dated back to 2011 and 2012, before Web Content Accessibility Guidelines 2.1, 2.2, HTML 5, or widely supported ARIA even existed, so the update was more than overdue. The most important structural change is that “recommendations” have been eliminated: everything on the list is now a requirement, some that were previously optional are now mandatory, and others have been removed outright as best practices that shouldn’t block approval.
Additionally, each requirement now follows a standard format with three sections: the basic principle, step-by-step testing instructions with clear pass/fail criteria, and links to relevant documentation. For theme authors who already have the label, the deadline to adapt to the new requirements is June 30, 2026. After that date, themes that don’t meet the new criteria will lose the label.
The WordPress Slack, the chat where the entire community communicates, is taking an important step toward internationalization: local communities worldwide will be able to integrate into the same space already used by the project’s contributor teams, without having to choose between their local community and global visibility. The initiative stems from a reality many already knew: much collaboration was happening in separate Slack instances, which made mutual discovery between communities and connection with Make WordPress teams difficult. Additionally, many of those local workspaces ran on Slack’s free plan, with limited message history and fewer available tools.
The goal is not to replace the work local communities are already doing, but to create better bridges. In practice, this means that meetup communities, regional WordCamps, and flagship events like WordCamp Europe, WordCamp US, and WordCamp Asia will have dedicated space within the global Slack. Some of these moves are already underway: WordCamp Asia and WordCamp US are already active within Make Slack, the Japanese community is in transition, and WordCamp Europe plans to join for the 2027 edition.
And finally, this podcast is distributed under a Creative Commons license as a derivative version of the podcast in Spanish; you can find all the links for more information, and the podcast in other languages, at WPpodcast .org.
Thanks for listening, and until the next episode!
Leave a Reply