Title: SolSEO
Author: Solkarra
Published: <strong>Ọ̀wàrà 4, 2026</strong>
Last modified: Ọ̀wàrà 10, 2026

---

Ṣàwárí àwọn plugin

![](https://ps.w.org/solseo/assets/banner-772x250.png?rev=3727295)

![](https://ps.w.org/solseo/assets/icon-256x256.png?rev=3727295)

# SolSEO

 Láti ọwọ́ [Solkarra](https://profiles.wordpress.org/hashyau/)

[Ṣe ìgbàsílẹ̀](https://downloads.wordpress.org/plugin/solseo.2.5.2.zip)

 * [Àwọn àlàyé](https://yor.wordpress.org/plugins/solseo/#description)
 * [Àwọn àgbéyẹ̀wò](https://yor.wordpress.org/plugins/solseo/#reviews)
 *  [Ìgbéwọlẹ̀](https://yor.wordpress.org/plugins/solseo/#installation)
 * [Ìdàgbàsókè](https://yor.wordpress.org/plugins/solseo/#developers)

 [Ìrànlọ́wọ́](https://wordpress.org/support/plugin/solseo/)

## Àpèjúwe

SolSEO does the SEO work every site needs and gets out of the way. Nothing is held
back behind a key, nothing phones home, and there are no notices on screens that
have nothing to do with it.

**A score you can act on**

Every post, page and product is scored out of a hundred from around thirty checks:
how the page is aimed at its keyword, titles and links, how the writing reads, and
what a shopping result needs. Each check says what to change. The score updates 
as you type, and the posts list gets a column you can edit in place, so forty titles
can be fixed without opening a page.

**Titles, meta and sitemaps**

Write a title and description per page, or set a template per post type with placeholders
in braces, such as {title} {sep} {sitename}. A pixel gauge shows when a title is
about to be cut off. XML sitemaps cover posts, pages, products, categories, tags
and author archives, list the images on each page, and are added to robots.txt.

**Structured data**

Organisation or person, the site, breadcrumbs, articles and products. A product 
with variations gets one offer per variation with its own price, SKU and stock state,
which is what sends a buyer to the right size.

**The rest**

 * A monthly crawl of up to three hundred of your own pages that turns up internal
   links going nowhere, each one naming the pages it is written on, and links arriving
   by way of two or more redirects. It only ever asks your own server for your own
   addresses
 * Redirects with 301, 302, 307 and 410, and a log of dead links that becomes a 
   redirect in one click
 * Internal link suggestions that write a real link and store a revision
 * An image report: files far bigger than the space they are drawn in, missing width
   and height, missing alt text, uploads no page uses
 * Six accessibility checks on your own content, each named by guideline
 * Open Graph and card tags, canonicals, noindex and nofollow, SEO fields on categories
   and tags, breadcrumbs
 * A robots.txt screen with ready made rule sets and a way back
 * Import from another SEO plugin, previewed first, originals left in place
 * Search Console figures for the page you are editing, on a read only permission

**Works with WooCommerce**

Product schema, product checks in the score, galleries in the sitemap, and price,
SKU and stock placeholders. Twelve checks from GoogleÌtumọ̀ Yorùbá: ’s product data
specification run against your products here rather than turning up in Merchant 
Centre three days later. WooCommerce is optional.

### External services

SolSEO does not contact anybody else on its own. Every service below is off until
you switch it on yourself, and each one stops the moment you switch it off, remove
its key or press Disconnect. This is the whole list.

Two things in this plugin ask your own server for your own pages, and neither is
one of these. The crawler on the Technical screen reads your published pages when
you press the button, and the tag check on the Connections screen reads your home
page when you press that button. Both go through the same piece of code, which refuses
any address that is not on your own site. Nothing about those pages leaves your 
server, so there is no third party to name here.

Some screens also carry a link to somebody elseÌtumọ̀ Yorùbá: ’s documentation, 
and a link is not a service. The robots.txt screen lists the AI crawlers by name
and links each one to what its operator publishes about it, at OpenAI, Anthropic,
Google, Common Crawl, Perplexity, ByteDance, Apple, Meta and Amazon. The Connections
screen links GoogleÌtumọ̀ Yorùbá: ’s pages about API keys, the analytics check links
each tag vendorÌtumọ̀ Yorùbá: ’s own installation guide, and the Australian checks
link the OAIC, the ACCC and the ATO. Nothing is fetched from any of them: the addresses
sit in a href and go nowhere until you click one, at which point your browser goes
there and this plugin is not involved. They are here so that what a screen tells
you about a crawler or a rule can be checked against the people who set it.

#### Google PageSpeed Insights, at www.googleapis.com

Off until you paste a Google API key on the Connections screen, and used only when
you press the button on the Speed screen.

Each check sends one address from your own site, plus your key, and Google sends
back how that page performs: a test it runs itself, and, if enough people have visited
the page recently, what those visits measured. Nothing else is sent, no check runs
on a schedule, and removing the key stops it.

The key is yours, from your own Google Cloud project. GoogleÌtumọ̀ Yorùbá: ’s terms
and privacy policy cover what they do with the address you send.

Terms: https://developers.google.com/terms
 Privacy policy: https://policies.google.
com/privacy

#### IndexNow, at api.indexnow.org

Off until you switch it on under SolSEO, Technical, Indexing.

When it is on, publishing or updating a page sends that pageÌtumọ̀ Yorùbá: ’s address,
once, so the search engines that take part know to come and look. One submission
reaches all of them: Bing, Yandex, Seznam and Naver read the same endpoint. Nothing
is sent for a draft, a private page, or a page you have asked to stay out of search
results.

Each submission carries your siteÌtumọ̀ Yorùbá: ’s host name, the page addresses,
the key described below and the address of the key file. Nothing else. The same 
screen also has a button that sends up to a thousand of your published addresses
in one go, which is useful the first time you switch it on; it only runs when you
press it, and only while IndexNow is on.

IndexNow also needs a key file at the root of your site, which this plugin creates
and serves for you. It holds a line of letters and numbers and nothing else, and
it is how the engines check the submission came from whoever runs the site.

IndexNow is run by its sponsors, Microsoft, Yandex and Seznam.cz.

Terms of use: https://www.indexnow.org/terms
 Privacy policy: https://privacy.microsoft.
com/privacystatement How it works: https://www.indexnow.org/documentation

#### Google Search Console, at searchconsole.googleapis.com

Off until you connect a Google account under SolSEO, Settings, Connections, and 
you can disconnect on the same screen.

Once connected, the plugin asks Google two things. At the moment you connect, and
again if you press the button to refresh the list, it asks which Search Console 
properties your Google account can see, so it can show you the list and match one
to this site. After that, when you open a post or a page in the editor, it asks 
what that one page did in Google over the last twenty eight days: clicks, impressions,
average position, and the search terms people used to reach it. The answer is kept
on your own server for six hours so that opening the same page twice does not ask
twice.

The permission asked for is read only. It is https://www.googleapis.com/auth/webmasters.
readonly, which is the narrowest one Search Console publishes: it cannot add a property,
remove one, or submit a sitemap. Nothing is ever written to your Google account.

What is sent is the address of the page you are editing and the address of the property
you matched it to. No page content, no customer data and nothing about your visitors.

None of what comes back is sent anywhere. The figures are shown in your own WordPress
admin and stored on your own server, and the plugin has no route that sends Search
Console data to SolSEO or to anybody else.

Terms: https://policies.google.com/terms
 Privacy policy: https://policies.google.
com/privacy

#### Connecting that Google account, at accounts.google.com, oauth2.googleapis.com and solseo.com.au

Off until you press Connect on that same screen. This is the one place in this plugin
that contacts solseo.com.au without a SolSEO account, and it is worth reading before
you press it.

Connecting to Google needs an application secret, and a secret that shipped inside
a plugin anybody can download is not a secret. So the plugin holds none. When you
press Connect, your browser goes to solseo.com.au, which sends you on to GoogleÌtumọ̀
Yorùbá: ’s own sign in and permission screen at accounts.google.com. When you say
yes, Google sends your browser back to solseo.com.au, which sends it straight back
to your site with a one-time code. Your site then asks solseo.com.au to turn that
code into a token at oauth2.googleapis.com. GoogleÌtumọ̀ Yorùbá: ’s tokens last 
an hour, so when the plugin next needs Search Console figures and finds the token
has run out, it asks solseo.com.au to refresh it the same way, sending the long-
lived refresh token Google issued.

What solseo.com.au sees is the one-time code, the refresh token when it is renewing
one, the token Google hands back, and the address of your site because your browser
came from it. It keeps none of them. There is no account, no row and no log line:
the code is exchanged and the answer is passed straight to your site, which is where
the token is stored, scrambled, in your own database.

If you would rather we were not in the middle at all, you do not have to be. Under“
Use my own Google app” on the same screen you can paste a client id and secret from
your own Google Cloud project, and then nothing in this handshake touches solseo.
com.au: your site talks to accounts.google.com and oauth2.googleapis.com directly.

Disconnecting deletes the stored token, tells Google to forget the permission, and
stops all of it in the same click. Deleting the plugin with “remove my data” ticked
under SolSEO, Tools, Data does the same: it tells Google to forget the permission,
once, before the stored token is deleted. You can also revoke it from your own Google
account at https://myaccount.google.com/permissions.

Google terms: https://policies.google.com/terms
 Google privacy policy: https://
policies.google.com/privacy SolSEO terms: https://solseo.com.au/terms SolSEO privacy
policy: https://solseo.com.au/privacy

#### SolSEO, at solseo.com.au

Off until you paste a pairing code on the Connect screen. This is separate from 
the Google handshake above, which uses solseo.com.au as a post box and needs no 
account.

Once paired, the plugin sends twice a day: the site address and time zone, the WordPress,
PHP, WooCommerce and plugin versions, the number of published items of each post
type SolSEO manages (posts, pages, products and any others you switched on), and
the permalink structure. In return it reads back which plan the account is on, a
summary of the sites on it, and what the accountÌtumọ̀ Yorùbá: ’s monitoring last
measured about this site from outside it, which is what the dashboard widget shows.
It only ever talks to solseo.com.au, whatever a reply from there says.

No page content, no customer data, no Search Console figures and nothing about your
visitors is sent. Disconnecting on the Connect screen removes the stored key and
stops it at once.

Service terms: https://solseo.com.au/terms
 Privacy policy: https://solseo.com.au/
privacy

## Àwọn àwòrán ìbòjú

[⌊The score and its checks in the editor sidebar, updating as you write.⌉⌊The score
and its checks in the editor sidebar, updating as you write.⌉[

The score and its checks in the editor sidebar, updating as you write.

[⌊The posts list, with the score, the keyword, the structured data and the link 
counts, and a pencil that edits titles and descriptions from the list.⌉⌊The posts
list, with the score, the keyword, the structured data and the link counts, and 
a pencil that edits titles and descriptions from the list.⌉[

The posts list, with the score, the keyword, the structured data and the link counts,
and a pencil that edits titles and descriptions from the list.

[⌊The SolSEO dashboard, with the average score, the pages worth looking at first
and the site checks.⌉⌊The SolSEO dashboard, with the average score, the pages worth
looking at first and the site checks.⌉[

The SolSEO dashboard, with the average score, the pages worth looking at first and
the site checks.

[⌊The robots.txt screen, showing what the site is serving and the sets of rules 
to choose from.⌉⌊The robots.txt screen, showing what the site is serving and the
sets of rules to choose from.⌉[

The robots.txt screen, showing what the site is serving and the sets of rules to
choose from.

[⌊Titles and meta templates, with the placeholders listed underneath.⌉⌊Titles and
meta templates, with the placeholders listed underneath.⌉[

Titles and meta templates, with the placeholders listed underneath.

[⌊The redirect manager and the log of addresses that found nothing.⌉⌊The redirect
manager and the log of addresses that found nothing.⌉[

The redirect manager and the log of addresses that found nothing.

## Àwọn ìdí

Plugin yìí pèsè 1 ìdí.

 *   Questions and answers A list of questions and answers, published as structured
   data so a search engine can read them.

## Ìgbéwọlẹ̀

 1. Install and activate the plugin.
 2. Open SolSEO in the admin menu. It works from the moment it is switched on, with
    no setup. If you would rather answer a few questions, SolSEO, Setup walks through
    five of them.
 3. Under Titles and Meta, set the organisation name and logo so the structured data
    is complete.
 4. If you are moving from another SEO plugin, go to SolSEO, Tools, Import, and check
    the preview before you copy anything.

## FAQ

### Does anything need an account?

No. Everything described above works with no account, no key and no network connection.

### Is there a paid version?

Yes, and nothing in this plugin is held back for it. SolSEO Pro is a separate add-
on that installs beside this one. It scores every phrase you are aiming at rather
than only the focus keyword, and it shows your tracked positions in the editor. 
There is a page about it under SolSEO, Upgrade, and at https://solseo.com.au/pro

### Will it clash with another SEO plugin?

Two plugins writing meta tags will produce two of everything. Import your fields
first, then switch the other one off.

### I imported by mistake. Is the old data gone?

No. The import copies. It never deletes or changes what the other plugin stored,
so switching that plugin back on puts things as they were.

### Will it stop AI crawlers using my writing?

There is a set of rules under SolSEO, Tools, robots.txt that names eight of them:
GPTBot, ClaudeBot, Google-Extended, CCBot, PerplexityBot, Bytespider, Applebot-Extended
and meta-externalagent. One button fills the box and nothing is saved until you 
press Save.

Three things are worth knowing, and the screen says all three. Two of those eight
do not fetch anything at all; they decide what may be done with pages an ordinary
search crawler already read, so saying no to them costs you nothing in search results.
One of them is a search crawler whose owner says it does not train on what it reads,
so blocking it takes your site out of that productÌtumọ̀ Yorùbá: ’s answers and 
saves no training use. And one of them has been seen fetching addresses it was told
to stay off. A robots.txt is a request. The well behaved crawlers follow it.

### Does it change the robots.txt file on my server?

No. WordPress generates robots.txt when there is no file, and that is what this 
adds to. If there is a real file on your server, nothing you type here is served,
and the screen says so and shows you what the file contains. It offers once to take
over: it copies the fileÌtumọ̀ Yorùbá: ’s lines into the box first so nothing is
lost, then renames the file to robots.txt.solseo-backup. It never deletes it, it
never does it on its own, and one button puts it back.

### Where did my robots.txt go?

If you took that offer, it is robots.txt.solseo-backup, in the same folder. “Put
it back as it was” on SolSEO, Tools, robots.txt returns it.

### Where is the sitemap?

At /sitemap.xml. It replaces the one WordPress generates.

### Can I change what a check expects?

The word count target can be filtered with `solseo_minimum_words`, and the whole
set of results can be filtered with `solseo_checks` before it is scored.

### Does deleting the plugin remove my SEO fields?

Not unless you ask it to, under SolSEO, Tools, Data.

## Àwọn àgbéyẹ̀wò

Kò sí àwọn àgbéyẹ̀wò fún plugin yìí.

## Àwọn Olùkópa & Olùgbéejáde

“SolSEO” jẹ́ ètò ìṣàmúlò orísun ṣíṣí sílẹ̀. Àwọn ènìyàn wọ̀nyí ti ṣe ìkópa sí plugin
yìí.

Àwọn Olùkópa

 *   [ Solkarra ](https://profiles.wordpress.org/hashyau/)

[Túmọ̀ “SolSEO” sí èdè rẹ.](https://translate.wordpress.org/projects/wp-plugins/solseo)

### Ṣe o nífẹ̀ẹ́ sí ìdàgbàsókè?

[Ṣàwárí koodu](https://plugins.trac.wordpress.org/browser/solseo/), ṣàyẹ̀wò [ibi ìpamọ́ SVN](https://plugins.svn.wordpress.org/solseo/),
tàbí ṣe àgbékalẹ̀ sí [àkọsílẹ̀ ìdàgbàsókè](https://plugins.trac.wordpress.org/log/solseo/)
nípasẹ̀ [RSS](https://plugins.trac.wordpress.org/log/solseo/?limit=100&mode=stop_on_copy&format=rss).

## Àkọsílẹ̀ àwọn àyípadà

#### 2.5.2

The sitemap now agrees with each pageÌtumọ̀ Yorùbá: ’s robots tag. Pages whose SEO
title or
 description was changed in the block editor are back in the sitemap. The
SolSEO panel saved an unticked “Keep this page out of search results” box as an 
empty value, and the sitemap left out every page carrying one. Updating clears those
values, so the pages return without being saved again, and a box unticked from now
on saves nothing. WooCommerceÌtumọ̀ Yorùbá: ’s cart, checkout and account pages 
are no longer listed: WooCommerce keeps them out of search, and Search Console reported
each one as “Submitted URL marked noindex”. The front page is left out while Titles
and Meta keeps it out of search. The shop page is still listed. Developers can leave
out a page another plugin hides with the solseo_sitemap_hidden_ids filter.

#### 2.5.1

The answer to the third wordpress.org review. Every value a form sends is now
 sanitised
by a WordPress function on the line that reads it: the robots.txt rules and the 
profile addresses through wp_strip_all_tags(), which keeps their lines and percent
encoded characters, before SolSEOÌtumọ̀ Yorùbá: ’s own checks run. The Titles, Writing
and Tools screens read each field with its own sanitiser rather than the whole form
at once. Redirects copied from another plugin have their targets cleaned as addresses,
and reading that pluginÌtumọ̀ Yorùbá: ’s table can no longer print a database error.

#### 2.5.0

Moving from another SEO plugin now brings its category and tag fields across
 with
the pages: titles, descriptions, canonicals, focus keywords, and whether a category
is kept out of search. A field SolSEO already holds on a category is never replaced.
The importÌtumọ̀ Yorùbá: ’s check covers categories too, so the offer to switch 
the other plugin off waits until they are all across. Redirects kept in the other
pluginÌtumọ̀ Yorùbá: ’s own table can be copied into SolSEOÌtumọ̀ Yorùbá: ’s, from
the Import tab or the last step of setup, with a list of what each one will do first.
Rules that matched part of an address come across as patterns that answer the same
addresses, one already here is left as it is, and one that cannot work here is named
with the reason. Both read the other pluginÌtumọ̀ Yorùbá: ’s data whether or not
it is still switched on. A pageÌtumọ̀ Yorùbá: ’s automatic description no longer
picks up a formÌtumọ̀ Yorùbá: ’s “please enable JavaScript” line, its labels or 
its button.

#### 2.4.0

WooCommerce products and categories can drop /product/ and /product-category/
 from
their addresses, so /product-category/string-wax/ becomes /string-wax/. The two 
switches are under Titles and Meta, on the Post types and Taxonomies tabs, and both
start off. A third removes parent categories from a subcategoryÌtumọ̀ Yorùbá: ’s
address. The old addresses send a permanent redirect to the new ones, and canonicals,
sitemaps and breadcrumbs print the new ones. A page keeps its address when a product
or category shares its name, and a product keeps it over a category, so switching
on never hides a page. A site moving from an SEO plugin that already removed these
keeps them removed: SolSEO reads that pluginÌtumọ̀ Yorùbá: ’s setting once, when
it is first activated. /sitemap_index.xml now sends a permanent redirect to /sitemap.
xml when nothing else on the site answers it.

#### 2.3.5

Saving one tab of Titles and Meta no longer switches off settings on another.
 Saving
the Social tab turned off structured data, the site search box, breadcrumbs and 
the search results noindex from the General tab, and saving General turned off Open
Graph and card tags from Social. Each tab now saves only the switches it shows. 
If you saved either tab on 2.3.4 or earlier, open both and check the ticks.

#### 2.3.4

The answer to the second wordpress.org review, plus the audit that followed it.

The sitemapÌtumọ̀ Yorùbá: ’s browser view carries no CSS at all, so nothing is printed
outside the enqueue system. Every request value is sanitised on the line that reads
it and every output is escaped where it is printed. The Google callback writes nothing
without a live handshake. The role map is enforced on every path that saves SEO 
fields, SQL lists use placeholders, the crawler verifies certificates, and deleting
the plugin with “remove my data” ticked now removes every table, option and stored
credential, and tells Google to forget the permission.

#### 2.3.3

No node in the structured data can point at a node that is not there. 2.3.2
 fixed
the one case Search Console had reported, the home page naming a breadcrumb list
it did not have. This release makes the rule general: after every builder and every
filter has run, any reference to an id that no node in the graph carries is removed.
An author reference to a person who has since been deleted, or a reference left 
behind by an add-on that replaced a node, can no longer reach the page as an empty
item that Google reports as a missing field. Pages whose references all resolve,
which is every page on a normal site, print exactly what they printed before.

#### 2.3.2

The home page no longer points at a breadcrumb trail it does not have. A trail
 
needs two crumbs and the front page has one, so no BreadcrumbList was written there,
but the WebPage node still named “#breadcrumb” as its breadcrumb. Google reads a
reference to an id nothing else carries as an empty BreadcrumbList, and Search Console
reported the home page as missing itemListElement. The page node now names a breadcrumb
only when the list is in the same graph. Every other page is unchanged.

#### 2.3.1

The WooCommerce shop page prints the title and description typed on it. The
 shop
is drawn at the address of a real page, and that page has the SolSEO meta box, but
the query calls the view an archive, so what was typed there was never read and 
the shop showed “Shop Ìtumọ̀ Yorùbá: – Sitename” whatever the editor said. The shop
archive now carries its page, the way the blog home already carried its, so title,
description, robots and social tags all come from the page. Every other archive 
is unchanged.

#### 2.3.0

A page title can no longer break the structured data. The JSON-LD block was
 written
with forward slashes left unescaped, so a title, description or product field holding
the six characters of a closing script tag ended the block early and everything 
after it in that field went onto the page as markup. The graph is now written so
that neither half of a tag can appear in it. Nothing is deleted: the words a person
typed are still in the graph, and a search engine reads the same page it always 
did.

The admin stylesheet no longer reaches WordPress’ own screens. Its colours were

declared on the document, so every screen that loaded it handed nine SolSEO values
to everything else drawn there, and it loaded on the Dashboard, the posts list, 
the editor and the term screen whether or not there was anything of ours on them.
The colours are now declared on this pluginÌtumọ̀ Yorùbá: ’s own elements, and the
stylesheet loads on a screen only when this plugin has drawn something on it: the
Dashboard only with one of its panels turned on, a posts list only for a post type
it manages, a term screen only for a taxonomy it manages.

The robots.txt rules ask WordPress where the admin is instead of assuming. A
 site
in a subdirectory, or one whose admin folder has been moved, was getting a rule 
for a folder that does not exist and an Allow line that did not cover admin-ajax.
php. Both lines now come from the address WordPress reports.

The sitemap stylesheet is a file. It used to be written into the transform,
 which
meant a browser fetched it again with every sitemap it drew.

Four links that had gone dead now point at pages that exist: ByteDanceÌtumọ̀ Yorùbá:’
s own
 page about its crawler, HotjarÌtumọ̀ Yorùbá: ’s tracking code guide, the 
ATO on tax invoices, and two ACCC pages behind the Australian checks. The readme
also now says, in the External services section, that the crawler and compliance
tables carry links to other peopleÌtumọ̀ Yorùbá: ’s documentation and that this 
plugin fetches none of them.

#### 2.2.0

The score in the editor now matches the audit. The editor analysis and the
 checks
that run on the SolSEO service had each grown their own copy of the same rules and
drifted apart, so a description could be written at a length the editor accepted
and then reported as too long by the same companyÌtumọ̀ Yorùbá: ’s check. The alt
text rule in the accessibility panel and the one in the SEO panel disagreed with
each other in the same editor. All of them now read one definition.

An empty alt attribute is read as an answer rather than as a missing one. A
 decorative
image is meant to carry an empty alt, and pages were being marked down for doing
it correctly.

#### 2.1.0

The dashboard widget is about this site and nothing else. It used to list every

site on the connected SolSEO account, so a site built for a client showed that client
the names and scores of every other site on the account. It now shows this one.

The panelÌtumọ̀ Yorùbá: ’s five sections are now also panels of their own: content
score, what
 needs work, site health and keywords, site monitoring, and pages worth
fixing. They start switched off. The Screen Options tab at the top of the Dashboard
turns any of them on, for you and not for everybody else on the site.

The widget also says a good deal more: how the pages are spread across the
 score
bands, how many have never been scored, how many have no search summary, how many
are hidden from search, and anything switched on that stops the site being found
at all. A connected site adds its health score and the change since the audit before,
its tracked keyword count, when it was last audited, and, where monitoring is on,
whether the site is answering, how long the certificate has left and whether the
sitemap works.

#### 2.0.2

Switching off the other SEO plugin after an import now switches off its paid
 add-
on at the same time. Before this, only the free plugin went off and the add-on kept
running with nothing underneath it, which took the whole site down until somebody
switched the first one back on.

The import screen also stops reloading itself. A finished run reloaded the page

once to show what it had unlocked, then the reloaded page saw the same finished 
run and reloaded again, over and over.

#### 2.0.0

Connect your Google Search Console account and see what each page actually did
 
in Google over the last twenty eight days, in the editor, beside the page you are
writing: clicks, impressions, average position, and the searches that brought them.
The permission asked for is read only. If you would rather solseo.com.au were not
in the middle of the connection, paste a client id and secret from your own Google
Cloud project and it will not be.

A product page no longer carries two Product descriptions. When SolSEOÌtumọ̀ Yorùbá:’
s
 structured data is on and it describes the product, WooCommerceÌtumọ̀ Yorùbá:’
s own copy is left out, so a search engine sees one product with one rating. A filter,
solseo_schema_replace_woo, keeps WooCommerceÌtumọ̀ Yorùbá: ’s instead. When another
SEO plugin is also publishing structured data, the Titles and Meta screen says so
in one line and asks you to turn one of them off. Add-ons can now hear when a pageÌtumọ̀
Yorùbá: ’s Search Console figures have been read.

#### 1.5.0

The free halves of six paid features: internal link suggestions you approve one

at a time, a warning when two pages go for the same phrase, a Merchant Centre feed
audit, six accessibility checks, six Australian business checks, and a verification
that your analytics tag fires once rather than twice or not at all. Also a crawl
of your own pages, the links on them that go nowhere, and what your images cost 
a visitor.

#### 1.3.1

The Upgrade screen described the add-on as adding two things over a list of
 three.
It no longer counts them in a sentence the list can outgrow.

#### 1.3.0

Seven SolSEO tests in WordPressÌtumọ̀ Yorùbá: ’s own Site Health screen, built from
what the
 plugin already knows and working with no account: whether search engines
are allowed in, readable addresses, the sitemap, structured data, pages with no 
description of their own, images with no alt text, and how many pages are kept out
of search results on purpose. A SolSEO section on the Info tab lists the version,
the post types scored, whether a sitemap is served and how the scoring stands, and
none of it leaves the site.

Two places an add-on can hook into, and the image description tool can now be
 pointed
at the images on one page.

#### 1.2.0

First release on wordpress.org.

A setup screen covering the five things a new site cannot guess. A panel in the 
block editor sidebar, opened from the score in the top right, in place of the box
below the content; the classic editor keeps its box. The posts list column now carries
the keyword, the structured data type and three link counts, and a pencil in its
header lets you edit titles and descriptions straight from the list.

The robots.txt screen shows what your site is actually serving, offers five ready
made sets of rules, records what the site served before this plugin arrived, and
can put it back. Anything with a per cent encoded character in it was being silently
stripped on save before now, so it is worth reopening that box.

Every bulk tool queues the whole job instead of asking you to press a button repeatedly.
The importer reads four more plugins, checks itself when it finishes, and then offers
to switch the other one off for you. A notice appears when another SEO plugin is
writing the same tags, on three screens only, dismissible for good.

Breadcrumbs have a live preview and the prefix setting finally has a field.

#### 1.0.0

Built, and never published.

## Àkójọpọ̀ Meta

 *  Ẹ̀yà **2.5.2**
 *  Ìgbàgbọ́hùn tó kẹ́yìn **ọjọ́ 4 sẹ́yìn**
 *  Àwọn ìgbéwọlẹ̀ tó ṣiṣẹ́ **Tó kéré sí 10**
 *  Ẹ̀yà WordPress ** 6.4 tàbí ju bẹ́ẹ̀ lọ **
 *  Dánwò dé **7.1.3**
 *  Ẹ̀yà PHP ** 7.4 tàbí ju bẹ́ẹ̀ lọ **
 *  Èdè
 * [English (US)](https://wordpress.org/plugins/solseo/)
 * Àwọn àmì
 * [meta](https://yor.wordpress.org/plugins/tags/meta/)[redirects](https://yor.wordpress.org/plugins/tags/redirects/)
   [schema](https://yor.wordpress.org/plugins/tags/schema/)[seo](https://yor.wordpress.org/plugins/tags/seo/)
   [sitemap](https://yor.wordpress.org/plugins/tags/sitemap/)
 *  [Ìwòye Tó Péye](https://yor.wordpress.org/plugins/solseo/advanced/)

## Àwọn ìbò

Kò sí ìwádìí tí a tíì fi ránṣẹ́.

[Your review](https://wordpress.org/support/plugin/solseo/reviews/#new-post)

[Wo gbogbo àwọn àgbéyẹ̀wò](https://wordpress.org/support/plugin/solseo/reviews/)

## Àwọn Olùkópa

 *   [ Solkarra ](https://profiles.wordpress.org/hashyau/)

## Ìrànlọ́wọ́

Nǹkan wà tí o fẹ́ sọ? Ṣé o nílò ìrànlọ́wọ́?

 [Wo àpéjọ ìrànlọ́wọ́](https://wordpress.org/support/plugin/solseo/)