I have modified en:User:Polygnotus/Scripts/Claude6.js to use Google Gemini's free as in beer and slightly better (at text in LMArena, anyway) gemini-2.5-flash-preview-05-20 model at en:User:Cramulator/GeminiProofreader.js. There is a hyperlink to Google's API key generation page in the key entry dialog box which anyone logged in to Google even without a Google Cloud account can use to generate a free tier key. I enabled Gemini's googleSearch and urlContext tools, so presumably it should be using its "grounding" capability to search for facts and sources, and actually examining linked sources (including PDFs etc.) when prudent. I have no experience with User/common.js gadgets so I hope others will improve on this approach.

I guess the main argument for it being less-friendly is that it does require a hypothetical completely-non-technical user to learn a new thing. We can assume that to do wiki admin stuff you're going to have to learn wikitext, but you could quite reasonably do everything on a wiki without ever needing to learn anything about Lua.

Having dived into this a bit, I am questioning the premise that having a non Lua template is "simpler", just because actually the templating language leaves a lot to be desired.

Change #1152382 merged by jenkins-bot:

[mediawiki/extensions/VisualEditor@master] Edit check: create editcheck-tone-shown tag

https://gerrit.wikimedia.org/r/1152382

I have a follow up question regarding row 6:
We have the following SLI to monitor model latency : 90% of all requests return a response within 1000 milliseconds
However in row 6 it states:

Proportion of edits that are published before the model is able to return an evaluation.

There is no way for the ML service to know when the edit is published. If we need to calculate this metric it will happen on the VisualEditor side. As I see the work to log this has already been implemented in https://phabricator.wikimedia.org/T388716#10872915
@ppelberg shall I add an 8th row to capture the latency SLI related to the LiftWing model I mention above?

Change #1154053 abandoned by Jdlrobson:

[mediawiki/extensions/VisualEditor@master] Don't pull in mediawiki.editfont.styles by default

https://gerrit.wikimedia.org/r/1154053

In T390248#10860617, @ppelberg wrote:

Next step

• @ppelberg to propose copy for card in light of offline discussion (May 21, 2025)

I've updated the task description to include the revised copy.

Change #1153717 abandoned by Jdlrobson:

[mediawiki/core@master] Remove mediawiki.editfont.styles from page load

Reason:

Loading of this is intentional.

https://gerrit.wikimedia.org/r/1153717

Got it, thanks for explaining the use case. Feel free to decline if this is intentional, I was worried it wasn't since I couldn't see any obvious usage. I'm not sure what the consequences of refactoring would be but trust your judgement there.

The current state is that the editfont styles are loaded as part of ext.visualEditor.tempWikitextEditorWidget, which is loaded with ext.visualEditor.desktopArticleTarget.init, which is loaded on every article page. The init module loads everything that it needs to bootstrap VisualEditor, and the temp wikitext widget is supposed to be a lightweight textarea that's functional for editing as quickly as possible if someone clicks "edit source" while the rest of the VE framework loads in the background.

Change #1153717 had a related patch set uploaded (by Jdlrobson; author: Jdlrobson):

[mediawiki/core@master] Remove mediawiki.editfont.styles from page load

https://gerrit.wikimedia.org/r/1153717

Change #1154053 had a related patch set uploaded (by Jdlrobson; author: Jdlrobson):

[mediawiki/extensions/VisualEditor@master] Don't pull in mediawiki.editfont.styles by default

https://gerrit.wikimedia.org/r/1154053