Skip to content

What Patcherly is (and is not)

Patcherly is an AI-powered automated program repair service for WordPress, WooCommerce, PHP, Python, and Node.js apps. It collects errors from your connectors, analyzes them with AI, proposes patches, then delivers them after approval - usually by applying on your site with a backup first and a post-patch check (with rollback when something goes wrong). A site can instead open a GitHub pull request (or open a PR and still apply live).

When you sign in, a quick onboarding checklist, optional guided tours, and a floating help entry help you learn the dashboard. Workspace owners can invite team members from Profile → Team (see Workspaces and team members).

What Patcherly does

Patcherly is an error-to-fix workflow:

  1. Connectors send errors as they occur, from the logs and queues you configure (Python, Node.js, PHP, or WordPress).
  2. Errors are analyzed with AI; a fix or patch is proposed on the error record in your workspace.
  3. After you (or policy) approve, the fix is delivered: the connector applies on your environment with a site backup before writes, or (Repo Pull Requests) Patcherly opens a GitHub PR for that site’s Connected repositories mode.
  4. For live applies, Patcherly runs checks after the patch so the app or site still works - for example a health request to your site URL, and (on supported plans) tests or deeper checks. If something looks wrong, rollback can run automatically or you can roll back manually. PR-only delivery waits for merge on GitHub instead.

Core mission: collect errors → analyze → patch → verify the change did not break things, with optional signals that the original issue may be resolved.

Key capabilities

  • Multi-language connectors (Python, Node.js, PHP, WordPress)
  • AI analysis with confidence scoring and an approval workflow (or auto-analysis / auto-apply on eligible plans)
  • Pre-apply backups on your server and rollback
  • Workspace metrics for error and fix activity (not a full external APM)
  • Patch Learning (Core and Pro): prior failed/rolled-back outcomes feed later AI analysis - in-product pattern recognition, not foundation-model training on your code (details)

What Patcherly is not

Patcherly is not a 24/7 uptime or liveness monitoring product. It does not watch your site every minute to tell you if it is “up” or “down” the way a dedicated monitoring or APM service does.

  • Audit, Metrics, and Usage show what happened in your workspace - they are not a replacement for a full-time external monitoring stack.
  • Optional agent heartbeat does not mean Patcherly stores continuous fleet health unless the product explicitly adds that later.

How it works (short)

  1. Install a connector and pair it (patcherly login or WordPress Pair connector).
  2. Errors are sent when the connector observes them.
  3. You click Analyze with AI (or auto-analysis starts).
  4. You review and approve the fix (or auto-apply when eligible).
  5. The fix is delivered (live apply with a backup, or a GitHub PR via Repo Pull Requests); live post-patch checks can trigger rollback.

Limits of “did the fix work?”

After a patch, checks are best-effort within the workflow:

  • Some problems only happen after specific user actions. Automated checks may confirm the site loads and tests pass, but cannot guarantee every interaction without manual QA.
  • Intermittent or environment-specific errors may not reproduce on every check.

Patcherly focuses on safe apply + post-patch verification, not exhaustive QA of every user journey.

Next steps