Suggestion Mode is a feature for the VisualEditor that proactively suggests actions that people can consider taking to improve Wikipedia articles. It is available as a Beta Feature (in your user-Preferences) on all Wikipedias.
This work is an effort to inspire people to publish constructive changes to Wikipedia that they might not otherwise while reducing toil involved with doing so.
In February 2026, Suggestion Mode was released as a Beta Feature at the English Wikipedia. In March 2026, it was released as a global Beta Feature. In the future we plan to make it available to newcomers (fewer than 100 edits).
Editing Wikipedia on the web often involves multiple browser tabs.
Newcomers are often unaware of the policies and guidelines that apply to the change(s) they are seeking to make and thus, unaware of the tabs they would benefit from opening.[1]
Experienced volunteers are often needing to consult a variety of pages (e.g. talk pages, watchlist, policy and guideline documentation, WikiProjects, noticeboards, etc.) in order to identify potential edits and in some cases, assemble the context needed to decide what (if any) action(s) to take.
This is particularly problematic on mobile devices where space and patience is limited.
There are a large number of newcomers who click "Edit" in mobile web and then don't attempt to save anything.[2][3]
With the above in mind, Suggestion Mode is meant to reduce the toil involved with:
Locating issues within Wikipedia
Assembling the context and tools needed to act
Why do mobile web users who click the edit button without publishing an edit?
This work is also grounded in research that found a large number of readers have been clicking the edit button because they're curious to try editing, but then closing the editor when they find no compelling edits to make.
In May and June 2026, Suggestion Mode was released through a controlled experiment at 15 Wikipedias ("test wikis").
At these test wikis, 50% of newcomers and Junior Contributors (defined as logged-in accounts who have published ≤100 cumulative edits) had Suggestion Mode enabled by default.
The other 50% of newcomers had Suggestion Mode disabled by default.
This experiment sought to learn whether presenting newcomers with the actionable edit suggestions Suggestion Mode offers would cause them to be more likely to publish a constructive edit.
Specifically, on mobile.
This experiment proved that while edit suggestions were available within 54.7% of the edit sessions newcomers in the test group initiated, newcomers only saw (read: clicked/tapped to view) a suggestion in 27% of said edit sessions.
Zooming out to the entire experiment, this means that newcomers in the test group saw a suggestion in 14.7% of the edit sessions they started.
Put another way, 85.3% of the treatment group effectively had an identical user experience to the control group, heavily diluting the feature's observable impact on overall experiment metrics.
As a result of the above, we did not see statistically significant changes – positive or negative – in any of the key metrics we sought to evaluate during the experiment.
We did not observe statistically significant differences (positive or negative) between the control and test groups in any of the metrics evaluated during the experiment:
Edit completion rate: proportion of qualified newcomer and Junior Contributor edit sessions [1] on a Wikipedia main namespace that are published.
[1] Qualified edit session = edit session in which the user spent at least ≥2 seconds in Visual Editor's ready state.
Second week retention: proportion of newcomers and Junior Contributors who return to publish at least one edit during their second week (day 7-14) after their initial constructive edit on a main namespace.
Constructive activation rate: proportion of newcomers making their first edit to an article in the main namespace within 24 hours of registration, and that edit not being reverted within 48 hours of being published.
Constructive edit rate: proportion of published edits in any Wikipedia main namespace by newcomers and Junior Contributors that are not reverted within 48 hours of being published.
The findings above can be explained by how relatively few people in the treatment group were exposed to the intervention (Suggestion Mode) which, in effect, diluted any change within the test group as a whole to the extent that there was no observable change within the broader population.
With the above said, when we narrowed our analysis down to the newcomers in the test group who saw a suggestion, we saw strong positive signals.
However, because filtering for users who actively clicked/tapped to view a suggestion inherently selects for a more motivated, higher-intent cohort, the scale of these results is likely inflated by selection bias.
There are two primary lessons from this first experiment:
Exposure: Suggestion Mode, as initially presented, couldn't move population-level metrics because most of the treatment group never encountered it. Newcomers in the treatment group saw a suggestion in 14.7% of the edit sessions they started.
Promising signs: despite the experiment not demonstrating any statistically significant movement in any of the key metrics we evaluated, when we "zoomed in" on the population of newcomers who did see a suggestion, the results were promising.
These findings land us squarely within "Scenario 4." outlined with the decision matrix we published before the experiment began: Suggestion Mode does not show a statistically significant impact on any KPI, secondary, or guardrail metric.
In line with the above, we are going to proceed by:
Publishing these results at all test Wikipedias and offering volunteers the choice to decide whether they'd like to enable Suggestion Mode as a default-on now.
Running a follow-up experiment that will show a visual indication to people who are logged in when edit suggestions are available within the articles they are reading to address the discoverability issue this experiment revealed. You can learn more about the design of the user experience and experiment here.
Volunteers can configure each Suggestion that is available within Suggestion Mode. This pattern is consistent with how Edit Checks are implemented. Existing configurations are listed at each wiki at their Special:EditChecks page, and can be adjusted by administrators at the linked json page.
Configurability will happen on a per project basis so that volunteers can ensure each Suggestion aligns with local policies and conventions and is integrated with existing moderation processes.
The aspects of a suggestion that are community configurable will vary on a case-by-case basis.
Each community can also create custom Suggestions. TextMatch is a locally-customizable feature that enables creating community-determined edit checks. It searches for predetermined queries and, when a match is found, displays the regular edit check dialog along with the corresponding message, title, and actions for that query.
If there are particular aspects of Suggestions that you think need to be configured on-wiki, we ask that you share what you are thinking in T327959 or on the talk page.
Suggestion Mode is related to the larger Edit check project in so far as the suggestions that appear within Suggestion Mode are implemented using the Edit Check platform/framework.
The project has been discussed in a few Community Conversations meetings, where volunteers expressed excitement about Suggestion Mode's potential to integrate community-developed tools into the core user experience.
If there is a project or resource you think we should be aware of, please add it here!
Duplicate reference finder: This Wscript detects and displays references with duplicate URLs. It also calculates how similar the visible text of those references is as a percentage. The script was created by a now banned user and is hosted by User:LuniZunie.
Template:Refideas editnotice: This template will cause an edit notice to be shown in an article when people have shared ideas for references on the talk page that could be used to improve it.
Gender Bias Detector – Scans Wikipedia biographies of women against 48 bias patterns across six categories and generates an editorial report with flagged passages and suggested rewrites.
Extension:InlineComments – an extension that adds the ability for viewers to add inline comments to a page, similar to inline comments in Confluence or annotations in tools like Google Docs.
Wikimonitor – a tool designed to monitor Wikimedia's RecentChanges stream in real-time. It empowers users to define custom filter rules using the Spring Expression Language (SpEL) to automatically flag specific edits, detect vandalism, or track specific patterns across Wikimedia projects.
Wikidata:ProVe – A tool that provides information about the quality of the references of Wikidata items, based on techniques such as large language models, triple verbalisation, and semantic similarity.
Meta:Cite Unseen: A user script that helps readers quickly perform an initial evaluation of the sources used in a given article. The script adds icons to identified sources that indicate various attributes of the source.
) in the edit toolbar. This confirms you've successfully enabled Suggestion Mode.
"Add a citation" example on the article w:en:Mango
"Convert citation" example on the article w:en:Gooseberry
"Revise tone" example on the article w:en:Melon
"External link" example on the article w:en:Atmosphere of Earth
"Missing caption" example on the article w:en:Guava
"Convert citation" (mobile) example on the article w:en:Gooseberry
"Missing caption" (mobile) example on the article w:en:Guava
Desktop feedback workflow
Mobile feedback workflow.png)
.png)
