Cursor can summarise these changes better than me, so here's the update according to our robot overlords...
What We Built Today
The Problem
Previously, to edit ProcessWire content, you had to either:
Log into the admin panel and use the web interface
Or ask the AI to read page content, but it couldn't write back
The Solution
We added Pull/Push sync — a Git-like workflow for ProcessWire content:
Pull a page from ProcessWire → saves it as an editable YAML file
Edit the YAML in your IDE (or have AI edit it)
Push the changes back → updates the live page
Key Features
Feature Description
Natural language -> Just say "Pull the About page" in Cursor chat
Mirrored folders -> Pages save to site/syncs/[page-path]/ matching your site structure
Readable YAML -> Content is human-editable, not database dumps
Readable dates -> Dates show as 2026-01-27 not Unix timestamps
Dry-run by default -> Push shows what would change before applying
Conflict detection -> Warns if someone else edited the page since you pulled
Revision tracking -> Each pull stores a hash to detect remote changes
What Works Now
Pull any page by path or ID
Edit text fields, titles, SEO fields, dates
Push changes back to ProcessWire
Re-pull to get latest version
What's Not Yet Implemented
Editing RepeaterMatrix content (the complex page builders)
Uploading/changing images and files
Bulk pull/push of multiple pages
Creating new pages
Example Workflow
You: "Pull the blog post about the Lake District"
AI: Finds and pulls it to site/syncs/news/posts/lake-district-walks/page.yaml (this mirrors the page path/slug)
You: Edit the YAML file + save
You: "Push the Lake District walks page"
AI: Shows preview, then applies changes to ProcessWire and updates the 'real' PW page