Changelog

Creating updates and training docs should not make you save first, close the dialog, reopen the record, and then hunt for the thing you actually meant to link. That flow has been cleaned up.

You can now choose related assets while creating updates and training docs. The create dialogs also look and behave more like the edit dialogs, so the fields are in a more predictable order and the linking section sits where it belongs: near the bottom with the rest of the record details.

Browser Connect now reads email templates from the same GHL email builder surface you see in the app. If an account had a pile of templates but PatchyHub only showed one failure, zero templates, or a suspicious missing-scope message, this is the fix for that nonsense.

Managed imports also got a cleaner handoff to Patchy's background brain. When you choose to let Patchy manage a new import, the Foundation module is included now. That means shared or cross-cutting assets placed in Foundation can still be analyzed instead of sitting in a weird unmanaged corner while everything else moves forward.

GoHighLevel changed how its browser app exposes the active session, which made older Browser Connect builds think you were not fully logged in even when you were staring directly at a logged-in GHL tab. Extremely helpful behavior from everyone involved.

PatchyHub GHL Connect is now on version 1.2.2. The connector can recover when the old browser-session path is missing by refreshing the GHL tab once and completing the secure handoff from the active browser session. If you have an older version installed, reinstall Browser Connect from the PatchyHub download page before importing.

Patchy got a lot less vague this week. Instead of opening with a pile of generic starter prompts, the empty chat now gives you one useful question and a tiny reason for asking it. The goal is simple: answer the thing that removes the most ambiguity, then move forward. Less questionnaire energy. More actual work getting done.

The system understanding flow also got deeper. Patchy can now work with the full claim ledger, reconcile confirmed system claims, and keep evidence language tighter when it is turning what it sees into documentation. Big imported systems got special attention too: asset tables and Patchy's counting tools no longer stop at the old 1,000-row ceiling, so large test accounts and real client monsters should report what is actually there instead of politely lying by omission.

The guide behind Patchy's documentation jobs is now repo-owned and generated into the active prompt path, which means the app, chat UI, tests, and prompt instructions are finally reading from the same source. Revolutionary concept, apparently.

Surveys have been a real PatchyHub asset type for months: importable, linkable, searchable, and deep-linked back into GHL. The silly part was visibility. Surveys were not marked as a primary asset tab, so unless you were looking through All assets or filtering manually, they looked like they had wandered off. They now have their own top-level tab beside the other main asset types.

Browser Connect also got a much less embarrassing email-template pass. Some GHL accounts were importing workflows, forms, funnels, surveys, trigger links, and basically everything except email templates. GHL has multiple template surfaces, builder routes, public-template behavior, and folder records mixed into the same neighborhood, because apparently one clean endpoint was too emotionally available. PatchyHub now checks the right email builder routes, skips folder records, and imports the actual templates instead of reporting zero like everything is fine.

The brain pipeline does a lot in the background now: setup, rebuilds, module creation, assignment, enrichment, scoring. That is good. The part where long-running jobs could look stalled, fail without enough context, or keep going after you wanted them cancelled was less good.

The job banner now has a real lifecycle. It tells you when work is running, when it looks delayed, when it finishes, and when it fails. Failed jobs show a details modal with enough context to understand what broke, and rebuilds can be retried without reconstructing the whole situation from memory. There is also a cancel flow now, which is useful when you accidentally ask the brain to chew through a pile of data you no longer care about.

Behind the scenes, the old v1 brain job path was removed so imports and rebuilds now hand off to the same brain setup flow. Fewer parallel pipelines. Fewer stale banners. Fewer places for reality to split. Always a plus.

Browser Connect now handles white-label GHL domains properly. If your agency is not using the default app.gohighlevel.com domain, the extension and PatchyHub handoff preserve that base URL so deep links keep opening in the right place. The popup also got cleaned up: it shows the selected sub-account instead of making you stare at raw URL fields like this is a networking exam.

Asset details got more useful too. PatchyHub now surfaces more imported metadata across custom fields, custom values, calendars, trigger links, funnels, pipelines, forms, surveys, email templates, conversation AI, memberships, objects, and snippets. Translation: less opening GHL just to answer basic questions like "what type is this field?", "where does this redirect?", or "what configuration came in with this asset?"

Re-imports are supposed to be boring. Add what is new, update what changed, preserve what you customized, and tell you what disappeared from the source file. PatchyHub now does more of that visibly.

The import flow now shows assets that were already in PatchyHub but were not found in the latest re-import, so you can decide whether they were removed from GHL or just absent from that export. Step-level errors are surfaced more clearly instead of getting swallowed in the machinery. Re-import link creation was tightened up so existing asset links do not collide with themselves. And if GHL hands over an item with no usable ID, PatchyHub refuses it instead of inventing a future duplicate problem for you.

There is also a Cancel Import button now. This should not have required a philosophical breakthrough, but here we are.

Billing got quieter and safer in the same window. Checkout and customer portal flows now resolve Stripe customers through the workspace instead of relying on email matching, and free-tier accounts no longer see a Manage Billing button they cannot use.

Patchy chat already learned how to use multiple tools in a single answer. This round made it better at understanding the room it is standing in. Responses now carry cleaner workspace and active-system context, so answers are less likely to confuse similarly named systems or drift into generic phrasing when you asked about the thing currently on screen.

Workspace health results now include the systems list Patchy needs for cross-system answers, and the module resolver got less jumpy when two names are close enough to look related. That means fewer "I think you meant..." moments where the assistant grabs the wrong system because a prefix matched and everyone pretended that was enough evidence.

The import entry point also got rearranged around how people actually use it. Browser Connect is now the recommended first card, token imports default into a brain-managed Unassigned destination, and the brain setup kicks off automatically after import so new data starts becoming useful without extra ritual clicking.

Answered every question on a module and curious if Patchy has anything new to say? There's now a small refresh button next to the module name in the center pane. Click it and Patchy re-examines the module on its next run (within 15 minutes). A 15-minute cooldown per module keeps you from hammering the same one repeatedly — if Patchy just looked, it'll tell you how long until it's worth looking again.