Collaborative editing is now built into Directus. Not an extension, not a plugin. It's in the core. That alone would make v11.15 a big release.
But we also shipped the AI Assistant to general availability with multi-provider support and real content awareness, plus a new deployment module that lets you trigger Vercel builds without leaving the Data Studio.
Let's get into it.
Breaking Changes
Visual Editor context for AI Assistant: The AI Assistant now accepts context from visual editor elements. If you're using @directus/visual-editing on your website, update to v1.2.0 or later to maintain compatibility.
Collaborative Editing: Built Into the Core
If you've used Directus with a team of any real size, you've hit the moment. Two people open the same record. Someone makes a change. The other person saves over it. Work disappears.
Collaborative editing was previously available as an extension. It worked, but having it live outside the core meant extra setup, extra maintenance, and limits on how deeply it could integrate with the rest of the platform.
As of v11.15, it's native. Multiple users can edit the same record at the same time, with live presence indicators showing who's working where.
Fields are locked at the field level while someone is editing them, so two people can't change the same field simultaneously. No conflicts, no last-save-wins.
Under the hood, it runs on WebSocket connections for real-time sync and Redis for distributed state coordination, which means it works across multi-instance deployments out of the box. And it plugs into your existing Directus permissions.
Users can only collaborate on records they already have access to. No new permission model to learn, no separate access controls to manage.
To get started, make sure WebSockets are configured on your instance. The feature is enabled by default (WEBSOCKETS_COLLAB_ENABLED), and you can tune timeout and cleanup intervals through environment variables if needed.
The bottom line: Your team can work on the same content at the same time without overwriting each other, and it's built right into Directus.
Check out the collaborative editing guide to enable it in your project settings.
AI Assistant: Context-Aware, Multi-Provider, and Production-Ready
When we shipped AI Chat in beta, it was a starting point. A conversational interface that could interact with your Directus instance. Useful, but it didn't know much about what you were actually working on.
v11.15 is where that changes.
The AI Assistant (formerly AI Chat) is now generally available, and it ships with three upgrades worth knowing about.
First, it supports multiple AI providers out of the box. OpenAI, Anthropic, Google Gemini, or any OpenAI-compatible provider like Ollama.
If your organization is already committed to a specific AI vendor for compliance or cost reasons, Directus works with what you have. Pick the provider and model that fits.
Second, the AI Assistant now understands context from your actual content. You can attach content items, predefined prompts, and visual editor elements directly to a conversation.
Instead of explaining what you're working on, you show it. Select a content item, stage a visual element from the page you're editing, and the assistant has the context it needs to give you specific, relevant responses rather than generic suggestions.
Third, the visual editor now connects to the AI Assistant with a dedicated sidebar. Work on a live preview of your page and pull the AI in without switching views. Select an element, ask the assistant to revise copy or suggest improvements, and see the result in context.
The live preview split pane also picked up display options, full-width mode with drag-to-expand, and quick access to the Visual Editor module.
Getting started is straightforward. Head to Settings, configure your AI provider with an API key, and the assistant becomes available throughout the Data Studio.
The bottom line: The AI Assistant knows what you're working on, works with the provider you choose, and is ready for production.
Check out the full AI Assistant guide to get set up.
Deployment Module: Ship From the Data Studio
Content is ready. It's reviewed, approved, published. Now someone needs to go trigger a deployment in Vercel. Or ping the dev team. Or remember which webhook URL to hit.
The new deployment module cuts that down to a button click. Connect your Vercel account, select the projects you want to manage, and trigger production or preview deployments directly from the Data Studio. You can monitor deployment runs and check logs without switching tools.
The setup takes about two minutes. Add your Vercel API token in the deployment settings, select which projects to connect, and you're ready to deploy. Your content team gets a clear, simple interface for something that used to require developer involvement or external tooling.
Vercel is the first supported provider, with more on the roadmap. Access is currently limited to admins, so you control who can trigger deployments.
The bottom line: Publish content and ship it to production without leaving Directus.
Check out the Vercel integration guide to connect your account.
Revision Comparison: Compare Against Previous Revisions
When you're reviewing content changes, you've always been able to compare the current version against a specific revision. But if you wanted to see what actually changed between two older revisions, there wasn't a great way to do that.
v11.15 adds a toggle that lets you compare any revision against the one right before it. It's a small change to the UI but a real improvement for reviewing content history, especially on teams where multiple people are editing across sessions. You can now trace exactly what changed and when, revision by revision.
This pairs well with the granular diff highlighting for WYSIWYG content changes that shipped in v11.14.1. Together, you get a clear, visual record of what changed between any two revisions, right down to the specific text edits.
Other Improvements
Relational fields on bar charts: You can now use relational fields on the x-axis of bar charts in Insights, opening up new ways to visualize your data relationships.
Permission-blocked field appearance: Fields you don't have permission to edit now show a non-editable appearance instead of looking disabled. It's a subtle distinction that makes it much clearer what you can and can't change.
ldapjs replaced with ldapts: The deprecated ldapjs library has been swapped for ldapts, its maintained successor.
Bug fixes: Fixed back button navigation on related items, markdown editor table options clipping, WYSIWYG rendering when a field is named "tooltip," caret jumping in template inputs, sidebar details not fetching on navigation, and inconsistent disabled states across interfaces.
Check out the full release notes on GitHub for the complete changelog.
Directus v11.15 is Available Now
Update your instances to get native collaborative editing, the production-ready AI Assistant, one-click Vercel deployments, and everything else in this release.
As always, back up your database before upgrading.