docs
/
AppEngine API

Sites

The site record — the home page, hosts and domains, logo and favicon, tracking and SEO — and the two fields whose shape is not what you would expect.

A site record is the container for everything that is true of a website rather than of one page: which page is the home page, what domain serves it, the logo, the favicon, analytics, and the site-wide SEO switches.

There is normally one per site, and an organization can have several.

s, d = req("GET", "/repository/find/site?perPage=10", token=t)
for r in d["data"]:
    dd = r["data"]
    print(dd["name"], dd.get("domain"), dd.get("hostName"))

The fields

{
  "name": "businessmade",
  "title": "BusinessMade",
  "description": "…",
  "keywords": "…",
  "homePage": "index",
  "hostName": "businessmade-businessmade.site.appmint.app",
  "domain": "businessmade.io",
  "logo": { "path": "…", "url": "https://…" },
  "favicon": "https://…/favicon.png?…",
  "faviconAssets": { "ico": "…", "png": "…" },
  "hideSiteHeader": false,
  "hideSiteFooter": false,
  "tracking": { "googleSiteVerification": "…", "bingSiteVerification": "…" },
  "seo": { "noIndex": false }
}

homePage — name, slug or id

"homePage": "index"

A page is found by its name, its slug (both case-insensitive) or its id — the same lookup every page address uses. / and /index both open it. If nothing matches, / is a 404 (the site's not-found screen). Studio's picker stores the page's name. How every other address is answered: Site page routing.

When the site sets publishedPagesOnly, pages are served from their published copies: an edit shows only after the page is published again (POST /repository/publish/page/:id).

s, d = req("GET", "/repository/find/page?perPage=100", token=t)
for r in d["data"]:
    print(r["name"], "|", r["data"].get("slug"))   # homePage may match either

favicon, logo and the icon routes

"logo":    { "path": "org/brand/logo.png", "url": "https://…" },   // Studio stores the picked file object
"favicon": "https://…/favicon.png?…"                                // Studio stores a string

The page never links to the stored URL. Every page's <head> names the site's own /favicon.ico, /favicon.png and /apple-touch-icon.png, and the share card falls back to /logo.png. Those routes read the site record per request and serve the bytes from the site's own domain (base-app lib/brand-asset.ts):

RouteFirst absolute URL among
/favicon.icofaviconAssets.ico, favicon, logo (none: 307 to the platform icon)
/favicon.pngfaviconAssets.png, favicon, logo (none: 404)
/apple-touch-icon.pngappleTouchIcon, logo, faviconAssets.png, favicon (none: 404)
/logo.pnglogo, faviconAssets.png, favicon (none: 404)

Each value may be a string, { url }, { signedUrl } or an array. An object favicon used to emit no icon link at all. That's fixed, but store a string anyway: it's what Studio and the generators write.

Each route sends the content type its name says. An SVG logo is served from /logo.png as SVG bytes labelled image/png, which breaks the share card and the apple icon. Use PNG or JPG.

Keep the generated pair under faviconAssets:

{ "data.favicon": "https://…/favicon.png?…",
  "data.faviconAssets": { "ico": "…", "png": "…", "path": "org/favicon.ico" } }

Verify on the site, never on the record. The <link rel="icon"> always says /favicon.ico; what matters is what that route serves:

curl -sI "https://your-site/favicon.png" | grep -iE '^HTTP|content-type|x-brand-asset'

A 200 with X-Brand-Asset: favicon.png means the record's icon is served. A 404 means no favicon, favicon pair or logo is set.

Page-level <link rel="icon"> tags in data.html are deliberately stripped by the renderer, so a page cannot set its own favicon. The site record is the only mechanism.

Generating the files: Set a logo and favicon.

Hosts and domains

FieldWhat it is
hostNameThe platform host — <site>-<org>.site.appmint.app. Set by POST /site/create-site and by Repair Site Domain. A site made with Studio's New Site (POST /site/add) has none; it keeps the same host in sharedDomain until a repair. A clone carries its source's.
domainYour custom domain
isDefaultThe site that answers any host no other site claims. One per org: setting it on one site clears it on the others.

Both serve the site, from the same renderer with cache-control: no-store; neither is cached by the platform. To connect, change or repair a custom domain, see Connect a custom domain.

Verify a deploy on hostName first. If the platform host shows your change and the custom domain does not, the domain is reaching something else: check its DNS with dig, or a cache in front of it on the owner's side (a Cloudflare proxy). A change that shows on neither is not rendering at all: a page saved in Build Studio renders data.screens, not data.html.

Tracking and SEO

"tracking": {
  "googleSiteVerification": "…",
  "bingSiteVerification": "…"
}

These come from the site record rather than a build-time variable, because one deployment serves many tenants and each claims its own property.

"seo": { "noIndex": true }

seo.noIndex is the site-wide kill switch and outranks any per-page value. It is the same flag robots.txt reads, so the two cannot disagree. It doesn't empty the sitemap: every page stays listed, each saying noindex. See Sitemap, robots and search indexing.

The verification codes are read from tracking only. A seo.googleSiteVerification or seo.bingSiteVerification is ignored.

Chrome switches

hideSiteHeader and hideSiteFooter suppress the platform's own header and footer. Set both if your pages carry their own — otherwise you get two headers, which usually reads as a CSS bug.

Writing to the record

s, rec = req("GET", f"/repository/get/site/{sk}", token=t)

req("POST", f"/repository/update-partial/site/{sk}", {
    "sk": sk,
    "version": rec["version"],
    "data.title": "New title",
}, token=t)

Dot-paths only. {"data": {...}} replaces the entire object — domain, logo, tracking and everything else you did not include.

Read the record before overwriting data.logo. A site may already have a designed logo, in which case you only want to set the favicon.