" And fallbacks.title is "<source>" Examples: | seo_title | template | title | source | | | %title% \| %site% | How we price a bathroom renovation \| Renew | title | | Bathroom costs | %title% \| %site% | Bathroom costs \| Renew | seo_title | | Bathroom costs | %site% - %title% | Renew - Bathroom costs | seo_title | | Bathroom costs | %title% | Bathroom costs | seo_title | | Bathroom costs | News: %title% | News: Bathroom costs | seo_title | Scenario Outline: An empty site_name leaves nothing in place of %site% and the title is trimmed Given the site 1 has site_name "" and title_template "<template>" And the article 12 has seo_title "Bathroom costs" When an anonymous caller sends GET /v1/article/12/seo Then title is "<title>" Examples: | template | title | | %title% \| %site% | Bathroom costs \| | | %site% %title% | Bathroom costs | | %title% | Bathroom costs | Scenario Outline: The description chain Given the article 12 has seo_description "<seo_description>", excerpt "<excerpt>", the body "<body>" and the site default_description "<default>" When an anonymous caller sends GET /v1/article/12/seo Then description is "<description>" And fallbacks.description is "<source>" Examples: | seo_description | excerpt | body | default | description | source | | Own words. | A teaser. | First one. Second. | Site. | Own words. | seo_description | | | A teaser. | First one. Second. | Site. | A teaser. | excerpt | | | | First one. Second. | Site. | First one. | body | | | | No terminator here | Site. | No terminator here | body | | | | <h2>Title</h2>What? Yes. | Site. | Title What? | body | | | | | Site. | Site. | site | | | | | | | none | Scenario Outline: The image chain Given the article 12 has social_image_id <social>, cover_image_id <cover> and cover_image_alt "<alt>", and the images 5, 6 and 7 exist When an anonymous caller sends GET /v1/article/12/seo Then image.url ends with "<path>", image.alt is "<image_alt>" and image.source is "<source>" And og.image equals image.url and twitter.image equals image.url And fallbacks.image is "<source>" Examples: | social | cover | alt | path | image_alt | source | | 6 | 5 | A field | /static/uploadedImages/6.png | | social | | null | 5 | A field | /static/uploadedImages/5.png | A field | cover | | null | null | | /static/uploadedImages/7.png | | site | Scenario: No image anywhere Given the article 12 has no cover and no social image and the site has no default social image When an anonymous caller sends GET /v1/article/12/seo Then image is {"url": "", "alt": "", "source": "none"} And og.image and twitter.image are "" And twitter.card is "summary" And fallbacks.image is "none" Scenario Outline: Canonical and robots Given the article 12 has canonical_url "<canonical>" and noindex <noindex>, and the site default_robots is "<robots>" When an anonymous caller sends GET /v1/article/12/seo Then canonical_url is "<resolved>" and og.url is "<resolved>" And fallbacks.canonical is "<source>" And robots is "<result>" Examples: | canonical | noindex | robots | resolved | source | result | | | false | index | https://www.renew.example/how-we-price-a-bathroom-renovation | site | index, follow | | | false | noindex | https://www.renew.example/how-we-price-a-bathroom-renovation | site | noindex, nofollow | | https://other.example/a | false | index | https://other.example/a | canonical_url | index, follow | | | true | index | https://www.renew.example/how-we-price-a-bathroom-renovation | site | noindex, nofollow | Scenario: Social overrides apply to og and twitter only Given the article 12 has social_title "Shared title", social_description "Shared text", seo_title "Meta title" and seo_description "Meta text" When an anonymous caller sends GET /v1/article/12/seo Then title is "Meta title | Renew" and description is "Meta text" And og.title and twitter.title are "Shared title" and og.description and twitter.description are "Shared text" Scenario: Social fields fall back to the resolved values independently Given the article 12 has social_title "Shared title" and no social_description When an anonymous caller sends GET /v1/article/12/seo Then og.title is "Shared title" and og.description is the resolved description Scenario: Authors carry the profile URL and only the social links in same_as Given the article 12 credits the author 2 "Marta Alves" with website "https://marta.example" and links [{"type": "github", "url": "https://github.com/marta"}, {"type": "x", "url": "https://x.com/marta"}] And the author 3 "Zoe Lambert" with no website and no links, credited after her When an anonymous caller sends GET /v1/article/12/seo Then authors is [{"name": "Marta Alves", "url": "https://marta.example", "same_as": ["https://github.com/marta", "https://x.com/marta"]}, {"name": "Zoe Lambert", "url": "", "same_as": []}] And the website "https://marta.example" is not in any same_as Scenario: An article without authors has an empty authors array When an anonymous caller sends GET /v1/article/12/seo Then authors is [] and json_ld has no author key Scenario: json_ld is a ready Article object Given the article 12 credits the author 2 "Marta Alves" as above and has the cover 5 with alt "A field" When an anonymous caller sends GET /v1/article/12/seo Then json_ld is {"@context": "https://schema.org", "@type": "Article", "headline": "How we price a bathroom renovation", "description": "We price by the square metre.", "image": ["https://www.renew.example/static/uploadedImages/5.png"], "author": [{"@type": "Person", "name": "Marta Alves", "url": "https://marta.example", "sameAs": ["https://github.com/marta", "https://x.com/marta"]}], "datePublished": "2026-09-01T08:00:00Z", "dateModified": "2026-09-10T09:30:00Z", "mainEntityOfPage": {"@type": "WebPage", "@id": "https://www.renew.example/how-we-price-a-bathroom-renovation"}} Scenario: The headline uses the meta title without the site template Given the article 12 has seo_title "Bathroom costs" When an anonymous caller sends GET /v1/article/12/seo Then json_ld.headline is "Bathroom costs" and title is "Bathroom costs | Renew" Scenario: The object never carries the body Given the body of the article 12 contains "square metre" When an anonymous caller sends GET /v1/article/12/seo Then no value of the response is the full body and the raw response contains no "" Scenario Outline: No format parameter is honoured When an anonymous caller sends GET /v1/article/12/seo<query> Then the status is 200 And the body equals the response without any query Examples: | query | | ?format=html | | ?format=json | | ?format=pdf | Scenario: Another site uses its own base URL, template and robots Given the article 12 moves to the site 2 with base_url "https://journal.renew.example", title_template "%title%" and default_robots "noindex" When an anonymous caller sends GET /v1/article/12/seo Then canonical_url starts with "https://journal.renew.example/" and robots is "noindex, nofollow" and title has no site suffix Rule: Drafts answer authenticated callers who can read the article Scenario: An online article is public When a caller without credentials sends GET /v1/article/12/seo Then the status is 200 Scenario Outline: A draft is visible to verified callers only Given the article 13 is a draft When GET /v1/article/13/seo is sent with <credentials> Then the status is <status> And the body is <body> Examples: | credentials | status | body | | the owner's jwt cookie | 200 | the resolved object | | a verified admin's jwt cookie | 200 | the resolved object | | a verified account with another role | 200 | the resolved object | | a valid API key | 200 | the resolved object | | no credentials | 404 | "Article not found" | | an unverified account's jwt cookie | 404 | "Article not found" | | a malformed jwt cookie | 404 | "Article not found" | Scenario: A draft has no publication date When the owner sends GET /v1/article/13/seo on a draft Then published_at is null and modified_at is the article's updated_at And json_ld has no datePublished Scenario: Unpublishing hides the object from the public again Given the article 12 is online When it is unpublished and an anonymous caller sends GET /v1/article/12/seo Then the status is 404 with body "Article not found" Scenario: A missing article is not found When any caller sends GET /v1/article/999/seo Then the status is 404 with body "Article not found" Scenario: A malformed id is refused When any caller sends GET /v1/article/abc/seo Then the status is 400 with body apperr.IdUint32ErrorMessage Scenario: The read is idempotent and writes nothing When the same request is sent twice Then both responses are identical and no repository write and no activity event happen Scenario: A repository failure answers 500 Given the article repository fails on its next call When a caller sends GET /v1/article/12/seo Then the status is 500 Rule: POST /v1/article/{id}/seo/check evaluates without saving Scenario: The check answers the documented shape Given the article 12 has the default state of the background When the owner sends POST /v1/article/12/seo/check with {} Then the status is 200 And the body has exactly the keys status, suggestions and slug And every suggestion has exactly the keys code, field and params And slug has exactly the keys available and suggestion Scenario Outline: Each code fires on its condition Given the article 12 is configured so that <condition> When the owner sends POST /v1/article/12/seo/check with {} Then suggestions contains the code "<code>" with field "<field>" and params <params> Examples: | condition | code | field | params | | the seo_title is 74 characters | title_long | seo_title | {"length": 74, "max": 60} | | the resolved title is empty after trimming | title_empty | seo_title | {} | | the resolved description is 41 characters | desc_short | seo_description | {"length": 41, "min": 70} | | the resolved description is 170 characters | desc_long | seo_description | {"length": 170, "max": 160} | | the cover is set and cover_image_alt is "" | cover_alt | cover_image_alt | {} | | cover_image_id is null | cover_none | cover_image_id | {} | | the slug is 76 characters | slug_long | slug | {"length": 76, "max": 75} | | no author is credited | no_au | author_ids | {} | | the body has 1000 characters of text and no h2 to h6 | no_h | body | {} | | the body has 3 images and one has an empty alt | body_alt | body | {"count": 1} | | the body has 40 characters of text | short | body | {"length": 40, "min": 300} | Scenario Outline: Soft ranges are inclusive at the edges When the owner sends POST /v1/article/12/seo/check with <body> Then the code <code> is <present> Examples: | body | code | present | | seo_title of 60 characters | title_long | absent | | seo_title of 61 characters | title_long | present | | seo_description of 70 characters | desc_short | absent | | seo_description of 69 characters | desc_short | present | | seo_description of 160 characters | desc_long | absent | | seo_description of 161 characters | desc_long | present | | slug of 75 characters | slug_long | absent | | slug of 76 characters | slug_long | present | Scenario Outline: The body rules have inclusive edges When the owner sends POST /v1/article/12/seo/check with a body of <text> characters of text, <headings> Then the code <code> is <present> Examples: | text | headings | code | present | | 999 | no heading | no_h | absent | | 1000 | no heading | no_h | present | | 1000 | one h2 | no_h | absent | | 1000 | one h6 | no_h | absent | | 1000 | only an h1 | no_h | present | | 300 | no heading | short | absent | | 299 | no heading | short | present | Scenario: Lengths are counted in runes When the owner sends POST /v1/article/12/seo/check with a seo_title of 60 characters that are 2 bytes each Then title_long is absent Scenario: The title range is measured on the meta title, not on the template output Given the site title_template adds 30 characters When the owner sends POST /v1/article/12/seo/check with a seo_title of 40 characters Then title_long is absent Scenario: Text length ignores tags When the owner sends POST /v1/article/12/seo/check with a body of 200 characters of text wrapped in 400 characters of markup Then the code short is present Scenario: A fully prepared article is ready Given the article 12 has a 45-character meta title, a 100-character description, a cover with alt text, a short slug, one author and a body with a heading, an image with alt text and 400 characters When the owner sends POST /v1/article/12/seo/check with {} Then status is "ready" and suggestions is [] Scenario: Body overrides replace saved values for the check only Given the article 12 has seo_title "Short" saved When the owner sends POST /v1/article/12/seo/check with {"seo_title": "<74 characters>"} Then suggestions contains title_long And the article 12 still has seo_title "Short" Scenario Outline: Overrides can also clear a saved value Given the article 12 has <saved> When the owner sends POST /v1/article/12/seo/check with <body> Then the code <code> is <present> Examples: | saved | body | code | present | | cover_image_id 5, empty alt | {"cover_image_id": null} | cover_alt | absent | | cover_image_id 5, empty alt | {"cover_image_alt": "A field"} | cover_alt | absent | | cover_image_alt "A field", cover_image_id 5 | {"cover_image_alt": ""} | cover_alt | present | | a 74-character seo_title and a 20-character title | {"seo_title": ""} | title_long | absent | Scenario: A body override feeds the body rules and the description fallback Given the article 12 has the saved body of the background and no seo_description and no excerpt When the owner sends POST /v1/article/12/seo/check with {"body": "Sentence one of exactly seventy characters, padded here to reach the minimum. Two."} Then the code desc_short is absent And the code short is present And the article 12 still has its saved body Scenario: A body override with images and headings When the owner sends POST /v1/article/12/seo/check with a body holding 1000 characters, no heading and an image without alt Then the codes no_h and body_alt are present Scenario: The check is non-persisting Given a snapshot of the article 12, its site and the activity log When the owner sends POST /v1/article/12/seo/check with {"slug": "other", "noindex": true, "excerpt": "x", "body": "x", "seo_dismissed": ["short"]} Then the article 12 is unchanged, updated_at included And no repository write, no activity event and no media usage change happen Scenario: Dismissed codes are filtered out Given the article 12 has seo_dismissed ["cover_none", "no_au"] and would raise cover_none, no_au and short When the owner sends POST /v1/article/12/seo/check with {} Then suggestions has only the code "short" Scenario: A dismissed override in the body filters the check When the owner sends POST /v1/article/12/seo/check with {"seo_dismissed": ["short", "no_au", "cover_none", "no_h"]} on an article that would raise those four only Then suggestions is [] and status is "ready" Scenario: Dismissing everything leaves ready, noindex still hides Given the article 12 has noindex true When the owner sends POST /v1/article/12/seo/check with {"seo_dismissed": [<all 11 codes>]} Then status is "hidden" and suggestions is [] Scenario Outline: The status Given the article 12 has <state> When the owner sends POST /v1/article/12/seo/check with {} Then status is "<status>" Examples: | state | status | | no suggestion and noindex false | ready | | two suggestions and noindex false | suggestions | | no suggestion and noindex true | hidden | | two suggestions and noindex true | hidden | Scenario: A noindex override gives hidden without saving When the owner sends POST /v1/article/12/seo/check with {"noindex": true} Then status is "hidden" and the article 12 still has noindex false Scenario: Suggestions come in a fixed order Given the article would raise short, title_long and no_au When the owner sends POST /v1/article/12/seo/check with {} Then the codes are in the order title_long, no_au, short Scenario Outline: Slug availability on the article's site Given the site 1 has the articles 12 with slug "how-we-price-a-bathroom-renovation" and 14 with slug "spring" and 15 with slug "spring-2", and the site 2 has an article with slug "autumn" When the owner sends POST /v1/article/12/seo/check with <body> Then slug is <slug> Examples: | body | slug | | {} | {"available": true, "suggestion": null} | | {"slug": "how-we-price-a-bathroom-renovation"} | {"available": true, "suggestion": null} | | {"slug": "fresh"} | {"available": true, "suggestion": null} | | {"slug": "autumn"} | {"available": true, "suggestion": null} | | {"slug": "spring"} | {"available": false, "suggestion": "spring-3"} | | {"slug": ""} | {"available": true, "suggestion": null} | Scenario: A regenerated empty slug can be taken Given another article of the site 1 already has the slug "how-we-price-a-bathroom-renovation" and the article 12 has slug "custom" When the owner sends POST /v1/article/12/seo/check with {"slug": ""} Then slug is {"available": false, "suggestion": "how-we-price-a-bathroom-renovation-2"} Scenario: The suggestion fits the slug limit Given another article of the site has a 100-character slug When the owner sends POST /v1/article/12/seo/check with that slug Then slug.suggestion is at most 100 characters and ends with "-2" Scenario: A taken slug is not a refusal When the owner sends POST /v1/article/12/seo/check with a taken slug Then the status is 200 and not 409 Scenario Outline: The same validation as PATCH applies When the owner sends POST /v1/article/12/seo/check with <field> <value> Then the status is <status> And the body is "<body>" Examples: | field | value | status | body | | seo_title | 121 characters | 400 | the validator error text | | seo_description | 321 characters | 400 | the validator error text | | excerpt | 301 characters | 400 | the validator error text | | cover_image_alt | 201 characters | 400 | the validator error text | | social_title | 121 characters | 400 | the validator error text | | social_description | 321 characters | 400 | the validator error text | | slug | "Spring" | 400 | invalid_slug | | slug | 101 characters | 400 | invalid_slug | | canonical_url | "http://example.com/a" | 422 | invalid_url | | cover_image_id | 999 (no such image) | 400 | image_not_found | | social_image_id | 999 (no such image) | 400 | image_not_found | | seo_dismissed | ["nope"] | 400 | the validator error text | Scenario Outline: Malformed bodies are refused When the owner sends POST /v1/article/12/seo/check with <body> Then the status is 400 And the body is apperr.BodyErrorMessage Examples: | body | | {not json | | {"slug": null} | | {"noindex": null} | | {"seo_dismissed": null} | | {"body": null} | | {"noindex": "yes"} | | {"colour": "x"} | Scenario: An empty body object is a valid check of the saved state When the owner sends POST /v1/article/12/seo/check with {} Then the status is 200 Scenario Outline: Site and author overrides are not accepted When the owner sends POST /v1/article/12/seo/check with <body> Then the status is 400 with body apperr.BodyErrorMessage Examples: | body | | {"site_id": 2} | | {"author_ids": [2]} | Scenario Outline: Anyone who can edit the article may check it When POST /v1/article/12/seo/check is sent with <credentials> Then the status is <status> Examples: | credentials | status | | the owner's jwt cookie | 200 | | a verified admin's jwt cookie | 200 | | a verified account with another role | 200 | | a valid API key | 200 | | an unverified account's jwt cookie | 403 | | no credentials | 401 | Scenario: A draft can be checked When the owner sends POST /v1/article/13/seo/check with {} on a draft Then the status is 200 Scenario: A missing article is not found When the owner sends POST /v1/article/999/seo/check with {} Then the status is 404 with body "Article not found" Scenario: A malformed id is refused When the owner sends POST /v1/article/abc/seo/check with {} Then the status is 400 with body apperr.IdUint32ErrorMessage Scenario: Publishing is never blocked by suggestions Given the check of the article 13 reports suggestions When the owner sends PATCH /v1/article/13/publish Then the status is 200 Rule: The article list carries the SEO status Scenario: Items carry seo_status and seo_suggestions Given the articles 1 (ready), 2 (2 suggestions), 3 (noindex true) and 4 (3 suggestions, one dismissed so 2 remain) When a verified account sends GET /v1/article/ Then the item 1 has seo_status "ready" and seo_suggestions 0 And the item 2 has seo_status "suggestions" and seo_suggestions 2 And the item 3 has seo_status "hidden" And the item 4 has seo_status "suggestions" and seo_suggestions 2 And every item has the existing fields, excerpt and slug, and no seo_dismissed key Scenario: A hidden article still reports its suggestion count Given the article 3 has noindex true and 2 non-dismissed suggestions When a verified account sends GET /v1/article/ Then the item 3 has seo_status "hidden" and seo_suggestions 2 Scenario: The list and the check agree Given the articles 1 to 4 above When the list is read and the check is run for each article with {} Then for each article seo_status equals the check status and seo_suggestions equals the number of check suggestions Scenario: Anonymous callers see online items with their status When a caller without credentials sends GET /v1/article/ Then each online item carries seo_status and seo_suggestions and no draft is listed Scenario Outline: The seo filter Given the articles 1 (ready), 2 (suggestions), 3 (hidden) and 4 (suggestions) When a verified account sends GET /v1/article/?seo=<value> Then the item ids are <ids> And total is <total> Examples: | value | ids | total | | ready | 1 | 1 | | suggestions | 4,2 | 2 | | hidden | 3 | 1 | Scenario: The filter composes with search, status, site, sort and paging Given 7 draft articles of the site "journal" titled "Tile ..." with suggestions, 3 online ones and 4 more on the site "renew" When a verified account sends GET /v1/article/?seo=suggestions&q=tile&status=draft&site=journal&sort=title:asc&size=5&page=2 Then the page holds the 2 remaining items in title order and total is 7 And pages 1 and 2 share no id Scenario: Counts stay unfiltered Given 4 online and 2 draft articles of which 3 have suggestions When a verified account sends GET /v1/article/?seo=suggestions Then counts is {"all": 6, "online": 4, "draft": 2} and total is 3 Scenario: The filter is applied before paging Given 30 articles of which 12 have suggestions When a verified account sends GET /v1/article/?seo=suggestions&size=5&page=3 Then the page holds 2 items and total is 12 Scenario: The filter respects visibility Given an offline article with suggestions When a caller without credentials sends GET /v1/article/?seo=suggestions Then the offline article is not listed and total ignores it Scenario: Dismissed codes are never counted in the filter Given an article whose only suggestion is dismissed When a verified account sends GET /v1/article/?seo=suggestions Then the article is not listed Scenario Outline: An invalid seo value is refused When a client sends GET /v1/article/?<query> Then the status is 400 And the body is apperr.BodyErrorMessage And the repository was not called Examples: | query | | seo=maybe | | seo= | | seo=READY | | seo=ready,hidden | | seo=1 | Scenario: Invalid format is still reported first When a client sends GET /v1/article/?seo=maybe&format=pdf Then the status is 400 and the body is "The format must be one of: markdown, html, json." Scenario: The existing list behaviour is unchanged without the filter Given the background articles When a verified account sends GET /v1/article/ Then total, page, size, counts, ordering and the other item fields are as before this slice Rule: GET /v1/site/{key}/sitemap lists published URLs Scenario: The sitemap lists the online articles of one site Given the site "renew" has the online articles 12 and 16, and a draft 13 And the site "journal" has the online article 20 When an anonymous caller sends GET /v1/site/renew/sitemap Then the status is 200 And the body is a JSON array of 2 items: {"url": "https://www.renew.example/how-we-price-a-bathroom-renovation", "modified_at": "2026-09-10T09:30:00Z", "noindex": false} and the item of the article 16 And the header X-Total-Count is "2" And each item has exactly the keys url, modified_at and noindex Scenario: Noindex articles are included with the flag Given the online article 16 has noindex true When an anonymous caller sends GET /v1/site/renew/sitemap Then the item of the article 16 has "noindex": true Scenario: The url uses the slug and not the canonical override Given the online article 12 has canonical_url "https://other.example/a" When an anonymous caller sends GET /v1/site/renew/sitemap Then the item url is "https://www.renew.example/how-we-price-a-bathroom-renovation" Scenario: Drafts are never listed When a caller with a verified jwt cookie sends GET /v1/site/renew/sitemap Then the draft 13 is not listed Scenario: An unpublished article leaves the sitemap Given the article 12 is unpublished When an anonymous caller sends GET /v1/site/renew/sitemap Then the article 12 is not listed and X-Total-Count decreases by 1 Scenario: The sitemap is public When a caller without credentials sends GET /v1/site/renew/sitemap Then the status is 200 Scenario: An empty site answers an empty array Given the site "renew" has only drafts When an anonymous caller sends GET /v1/site/renew/sitemap Then the status is 200, the body is [] and X-Total-Count is "0" Scenario Outline: Paging and the total header Given the site "renew" has <articles> online articles When an anonymous caller sends GET /v1/site/renew/sitemap<query> Then the body holds <count> items And X-Total-Count is "<articles>" Examples: | articles | query | count | | 250 | | 100 | | 250 | ?size=25 | 25 | | 250 | ?page=3 | 50 | | 250 | ?page=2&size=100 | 100 | | 250 | ?size=500 | 250 | | 250 | ?page=99&size=25 | 0 | Scenario: Pages partition the sitemap in a stable order Given 60 online articles When pages 1, 2 and 3 are requested with size 25 Then they hold 25, 25 and 10 items in id ascending order with no repeat Scenario Outline: Invalid paging answers 400 When a client sends GET /v1/site/renew/sitemap?<query> Then the status is 400 And the body is apperr.BodyErrorMessage Examples: | query | | page=0 | | page=-1 | | page=abc | | page= | | size=0 | | size=10 | | size=1000 | | size=abc | Scenario: An unknown key is not found When a client sends GET /v1/site/nope/sitemap Then the status is 404 with body "site_not_found" Scenario: The key is not an id When a client sends GET /v1/site/1/sitemap on a site whose key is "renew" Then the status is 404 Scenario: The route does not disturb the owner-only site routes When an admin sends PATCH /v1/site/2 with {"site_name": "x"} Then the status is 403 with body "Only the owner can do this." as before Scenario: The browser can read the total header When a cross-origin client reads the response of GET /v1/site/renew/sitemap with CORS configured Then Access-Control-Expose-Headers lists X-Total-Count and still lists Link Scenario: A repository failure answers 500 Given the article repository fails on its next call When a client sends GET /v1/site/renew/sitemap Then the status is 500 Rule: Layering and documentation Scenario: Layering is respected Then use cases import no persistence, mailer, models, gorm, net/http or chi And the resolver, the suggestion rules and the sitemap logic live in use cases or domain packages And the domain packages import no GORM or HTTP And handlers import no persistence and no repository And only api/dependecyInjection.go builds the new use cases and api/router.go wires them And no comment is added anywhere Scenario: The docs describe the new surface Then docs/api.md documents GET /v1/article/{id}/seo, POST /v1/article/{id}/seo/check, the seo query parameter, the seo_status and seo_suggestions item fields, GET /v1/site/{key}/sitemap, its X-Total-Count header and the CORS exposure of that header And docs/api/articles.md documents the resolved object, the fallback chain, the suggestion codes with their soft ranges and the draft visibility rule And docs/release-notes.md announces the change And web/src/data/apiCatalogue.ts lists the three new endpoints and the new list parameter and fields, and its spec and web/src/views/IntegrationView.spec.ts counts are updated

Feature: SEO output backend, resolved SEO object, non-persisting check, list status and sitemap As a frontend developer or the RenewCMS back-office editor I want the resolved SEO data of an article, a non-persisting check, an SEO status on the article list and a sitemap per site So that a frontend can render meta tags and a sitemap without reimplementing the fallback chain Background: Given the spec of record is docs/design/seo-slices.md (slice 2), docs/design/change-list.md ("SEO round"), docs/design/screens/editor.md, docs/design/screens/articles.md and docs/design/screens/integration.md And this slice is backend only: no SPA screen and no change to docs/design And it reuses the article and site SEO fields of features/seo_data_backend.feature and adds no column and no migration And RenewCMS renders no page and injects no meta tag: it only returns data And one rule set (a single use-case-level function over the article, its site, its authors and its body) feeds the list seo_status and seo_suggestions, the check endpoint and the suggestion-related parts of the resolver And the suggestion codes are title_long, title_empty, desc_short, desc_long, cover_alt, cover_none, slug_long, no_au, no_h, body_alt and short And the soft ranges are: meta title 30 to 60 characters, description 70 to 160, slug "very long" above 75; they only produce suggestions and never block saving or publishing And codes listed in the article's seo_dismissed are never reported and never counted And the status of an article is "hidden" when noindex is true, otherwise "suggestions" when at least one non-dismissed suggestion remains, otherwise "ready" And a caller is "verified" when the jwt cookie (or an API key) resolves to a user with is_verified true, and "public" otherwise, exactly as for GET /v1/article/{id} And the fallback chain is: title = seo_title, then the article title, with the site title_template applied last; description = seo_description, then excerpt, then the first sentence of the body, then the site default_description; image = social_image_id, then cover_image_id, then the site default_social_image_id; canonical = canonical_url, then site base_url + "/" + slug; robots = "noindex, nofollow" when noindex is true, otherwise from the site default_robots; og and twitter title and description = social_title and social_description, then the resolved values And the sample site 1 is key "renew", base_url "https://www.renew.example", site_name "Renew", title_template "%title% | %site%", default_description "Renovation notes from the Renew team.", default_robots "index", locale "en" and default_social_image_id 7 (path /static/uploadedImages/7.png) And the sample article 12 is online on the site 1, titled "How we price a bathroom renovation", slug "how-we-price-a-bathroom-renovation", created 2026-09-01T08:00:00Z, updated 2026-09-10T09:30:00Z, with no SEO field set and the body "

We price by the square metre. Tiles cost extra.

" And error responses are plain text from http.Error: 404 "Article not found" (apperr.ArticleNotFoundErrorMessage), 404 "site_not_found" (apperr.SiteNotFoundErrorMessage), 400 apperr.IdUint32ErrorMessage for a malformed id, 400 apperr.BodyErrorMessage for a malformed body or paging parameter And request validation goes through validate tags and validator.Validate.Struct And handlers reach articles, sites, authors and images only through use cases and domain interfaces, and no comment is added anywhere Rule: GET /v1/article/{id}/seo returns the resolved object Scenario: The object has the documented shape When an anonymous caller sends GET /v1/article/12/seo Then the status is 200 And the body has exactly the keys title, description, canonical_url, robots, og, twitter, image, authors, published_at, modified_at, fallbacks and json_ld And og has exactly the keys type, title, description, image, image_alt, url, site_name and locale And twitter has exactly the keys card, title, description, image and image_alt And image has exactly the keys url, alt and source And fallbacks has exactly the keys title, description, image and canonical And the body has no body, no id and no seo_dismissed key Scenario: A bare article resolves every output through the fallbacks When an anonymous caller sends GET /v1/article/12/seo Then title is "How we price a bathroom renovation | Renew" And description is "We price by the square metre." And canonical_url is "https://www.renew.example/how-we-price-a-bathroom-renovation" And robots is "index, follow" And image is {"url": "https://www.renew.example/static/uploadedImages/7.png", "alt": "", "source": "site"} And og is {"type": "article", "title": "How we price a bathroom renovation | Renew", "description": "We price by the square metre.", "image": "https://www.renew.example/static/uploadedImages/7.png", "image_alt": "", "url": "https://www.renew.example/how-we-price-a-bathroom-renovation", "site_name": "Renew", "locale": "en"} And twitter is {"card": "summary_large_image", "title": "How we price a bathroom renovation | Renew", "description": "We price by the square metre.", "image": "https://www.renew.example/static/uploadedImages/7.png", "image_alt": ""} And fallbacks is {"title": "title", "description": "body", "image": "site", "canonical": "site"} And published_at is "2026-09-01T08:00:00Z" and modified_at is "2026-09-10T09:30:00Z" Scenario Outline: The title chain and the template Given the article 12 has seo_title "" and the site title_template is "