Marketing site, SEO and AI SEO: pushizo.com launch steps
Pushizo's marketing site (apps/web, https://pushizo.com), developer documentation (apps/docs-site, https://docs.pushizo.com) and customer dashboard (apps/dashboard, https://app.pushizo.com; API https://api.pushizo.com)…
Pushizo's marketing site (apps/web, https://pushizo.com), developer documentation (apps/docs-site,
https://docs.pushizo.com) and customer dashboard (apps/dashboard, https://app.pushizo.com; API https://api.pushizo.com)
are ready for search engines and AI assistants. Builds made without SITE_URL / DOCS_SITE_URL are generated as
noindex. This guide lists, in order, the steps needed to launch pushizo.com.
Design decisions: ADR-0034.
1. Surfaces and environment variables
| Surface | Example domain | Variable | When it is read |
|---|---|---|---|
Marketing site (apps/web, port 3200) | pushizo.com | SITE_URL | Build time (image build argument) |
Documentation (apps/docs-site, port 3300) | docs.pushizo.com | DOCS_SITE_URL | Build time (image build argument) |
Customer dashboard (apps/dashboard, port 3100) | app.pushizo.com | APP_URL | Runtime |
Other marketing site variables (all read at build time):
| Variable | Default | Description |
|---|---|---|
WEB_DASHBOARD_URL | APP_URL, otherwise http://localhost:3100 | Root of the "Start for free" and "Sign in" links (/register, /login). |
WEB_DOCS_URL | DOCS_SITE_URL, otherwise http://localhost:3300 | Root of documentation links. |
WEB_CONTACT_EMAIL | empty | Contact email on the About page and in the Organization JSON-LD. If empty, no email is published. |
WEB_BILLING_ENABLED | false | true: the payment provider is active; paid plans lead to sign-up and the SoftwareApplication JSON-LD offers is published. false: paid plans show "Contact sales" (mailto: WEB_CONTACT_EMAIL, or support@pushizo.com if empty), no offers. |
WEB_COMPANY_NAME, WEB_COMPANY_ADDRESS, WEB_COMPANY_TAX_ID, WEB_COMPANY_MERSIS | empty | Company imprint: in the footer and on the /legal/imprint page. If the company name is empty, no imprint is published and the page returns 404; the MERSİS number (Turkish central registry number) has 16 digits. |
WEB_AI_SEARCH_CRAWLERS | allow | Answer/search AI agents (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot…). |
WEB_AI_TRAINING_CRAWLERS | allow | Model training crawlers (GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, CCBot…). |
INDEXNOW_KEY | empty | IndexNow key (8–128 letters/digits/hyphens). If set, /indexnow-key.txt is published. |
When SITE_URL is empty, the build prints the following warning and the site is generated with noindex, nofollow +
Disallow: / in robots.txt: [web] UYARI: SITE_URL tanımlı değil…. The same rule applies to the documentation for
DOCS_SITE_URL, and to the dashboard when APP_URL is not https.
2. DNS and TLS
- Create the root domain record for the marketing site (
A/AAAAor your provider'sALIAS/ANAMErecord) andA/CNAMErecords fordocs.andapp.. If you usewww., set up a permanent (301/308) redirect from one to the other; serving content on both addresses creates duplicate content.SITE_URLmust be the single chosen address. - Compose (single server): set
WEB_HOST,DOCS_HOSTandAPP_HOSTin.env.prod; for the initial certificates run./infra/scripts/certbot-init.sh(all domains, including WEB_HOST), thendocker compose --env-file .env.prod -f infra/compose/docker-compose.prod.yml up -d web docs nginx. - Kubernetes (Helm): set
web.enabled=trueandingress.hosts.web=<domain>; the certificate is added to the ingress TLS via cert-manager. Thewebsection invalues.yamlcontains resource and health check settings.
3. Release build
The release workflow (.github/workflows/release.yml) builds the images using GitHub repository variables. Define
the following under Settings → Secrets and variables → Actions → Variables:
| Repository variable | Value |
|---|---|
PROD_SITE_URL | https://pushizo.com |
PROD_DOCS_URL | https://docs.pushizo.com |
PROD_APP_URL | https://app.pushizo.com (may already be defined) |
PROD_INDEXNOW_KEY | optional IndexNow key |
Manual build example:
docker build -f infra/docker/web.Dockerfile \
--build-arg SITE_URL=https://pushizo.com \
--build-arg WEB_DASHBOARD_URL=https://app.pushizo.com \
--build-arg WEB_DOCS_URL=https://docs.pushizo.com \
-t ghcr.io/exec0de/pushizo-web:local .
docker build -f infra/docker/docs-site.Dockerfile \
--build-arg DOCS_SITE_URL=https://docs.pushizo.com \
--build-arg SITE_URL=https://pushizo.com \
-t ghcr.io/exec0de/pushizo-docs-site:local .Staging: the release image is the same in staging and production; block crawlers on the staging domain at the
edge (add add_header X-Robots-Tag "noindex, nofollow" always; to the nginx server block, or use basic auth). Because
canonical URLs point to the production domain, staging pages won't be indexed as duplicates, but it's still
recommended that they aren't crawled.
4. Post-launch verification
curl -s https://pushizo.com/robots.txt # Allow + Sitemap line, AI crawler groups
curl -s https://pushizo.com/sitemap.xml # index listing the 22 language sitemaps
curl -s https://pushizo.com/sitemap/de.xml | head # absolute https URLs, xhtml:link alternates
curl -sI https://pushizo.com/ | grep -i x-robots-tag # there should be NO output (noindex must be gone)
curl -s https://pushizo.com/llms.txt | head
curl -s https://docs.pushizo.com/llms.txt | head
curl -s https://app.pushizo.com/robots.txt # only /login and /register allowed- Use Google's Rich Results Test and the Schema Markup Validator to validate the home page (Organization,
WebSite, SoftwareApplication),
/faq(FAQPage) and a documentation page (TechArticle). - Use PageSpeed Insights / Lighthouse to check that performance, accessibility and SEO scores are ≥ 95 on both mobile and desktop.
- Social cards: check that the
https://<domain>/og/tr/homeimage loads (Arabic, Japanese, Korean, Chinese, Hindi and Thai pages use the English image because the embedded font does not cover those scripts) and that the title appears correctly in LinkedIn Post Inspector / the X card preview.
5. Google Search Console
- Add a Domain property (DNS TXT verification; covers all subdomains and both http/https versions).
- Sitemaps: submit
https://pushizo.com/sitemap.xmlandhttps://docs.pushizo.com/sitemap.xml. The site'ssitemap.xmlis an index listing the per-language sitemaps (/sitemap/<lang>.xml, 22 languages; ADR-0084); Search Console reads the sitemaps in the index on its own. - Use URL Inspection to "Request indexing" for the home page,
/en,/de,/pricingand/features/push. - After a few days, the Pages report should not show "Excluded by 'noindex' tag" (if it does, the image was
built without
SITE_URL). No separate International targeting setting is needed; hreflang tags are on the page. - Only
/loginand/registerare indexed on the dashboard; other dashboard URLs appear as "Blocked by robots.txt", which is expected.
6. Bing Webmaster Tools and IndexNow
-
In Bing Webmaster Tools, use Import from Google Search Console to bring in the property and sitemaps (or verify with DNS/an XML file and submit the sitemaps manually).
-
IndexNow: generate a random key (
openssl rand -hex 16), define it asPROD_INDEXNOW_KEYand rebuild the site;https://pushizo.com/indexnow-key.txtmust return the key. After each release:SITE_URL=https://pushizo.com INDEXNOW_KEY=<anahtar> npm run indexnow --workspace @pushizo/websiteThe script submits the URLs of every language sitemap in the sitemap index to
api.indexnow.org(Bing, Yandex and other IndexNow partners).
7. Legal pages
The Privacy Policy, Terms of Use, KVKK (Turkish data protection law) Privacy Notice, Cookie Policy, Data Processing
Agreement (dpa), Subprocessors (subprocessors) and Acceptable Use Policy (aup) are published from the following
files (both languages are required; if one is missing, the page is not generated). The imprint (/legal/imprint)
does not come from these files; it is generated when WEB_COMPANY_* is set:
apps/web/content/legal/privacy.tr.md apps/web/content/legal/privacy.en.md
apps/web/content/legal/terms.tr.md apps/web/content/legal/terms.en.md
apps/web/content/legal/kvkk.tr.md apps/web/content/legal/kvkk.en.md
apps/web/content/legal/cookies.tr.md apps/web/content/legal/cookies.en.md
apps/web/content/legal/dpa.tr.md apps/web/content/legal/dpa.en.md
apps/web/content/legal/subprocessors.tr.md apps/web/content/legal/subprocessors.en.md
apps/web/content/legal/aup.tr.md apps/web/content/legal/aup.en.mdMissing files are listed by npm run check:legal --workspace @pushizo/website (it only warns and does not stop the
build).
Optional front matter ---\nupdated: 2026-10-16\n--- is shown on the page as "Last updated". Once the files are added,
pages are generated at /legal/<name> and /en/legal/<name> and are automatically included in the sitemap, the
footer and llms.txt. Legal texts are published only in Turkish and English (ADR-0084 §5): in the other 20 languages the
footer links to the English text and shows a note, and /<lang>/legal/<name> redirects to /en/legal/<name>.
Because the site doesn't use cookies (the theme preference is kept in browser storage), there is no cookie consent banner; if an analytics tool is added, the cookie policy and consent mechanism must be re-evaluated.
For vulnerability reports, /.well-known/security.txt (RFC 9116; Contact: mailto:security@pushizo.com, Expires
one year after build time) is generated during the build: the site must be rebuilt and redeployed before it expires.
The abuse reporting address (abuse@pushizo.com) is published on the About page.
8. AI visibility (AI SEO)
llms.txt/llms-full.txtare published on the marketing site and in the documentation; since their content is generated from the site copy and the plan matrix, they need no separate maintenance.- AI crawlers are allowed by default. To block model training crawlers, rebuild with
WEB_AI_TRAINING_CRAWLERS=disallow; to block answer/search agents, rebuild withWEB_AI_SEARCH_CRAWLERS=disallow. - When updating copy, keep definitional sentences short, factual and self-contained; don't write prices and limits into the text (placeholders are filled from the plan matrix).
9. Local development
npm run dev --workspace @pushizo/website # http://localhost:3200 (Turkish /, English /en)
npx turbo run build --filter=@pushizo/website # without SITE_URL: a noindex warning is expected
npx playwright test tests/e2e/web-site.spec.ts # pages (22 languages), language switcher, SEO tags, axeDMARC and TLS-RPT monitoring
Receiving DMARC aggregate reports and TLS-RPT reports for the platform domain (Postal HTTP endpoint or IMAP), parsing, unauthorized source alerts, policy recommendations and one-click p= changes with Cloudflare.
Runbook: Incident management and status page
Sources: Alertmanager (#pushizo-alerts, #pushizo-oncall, email), Grafana dashboards (infra/monitoring/grafana/dashboards), Loki logs (tracing by request_id), Sentry. Status page: https://status.<domain> →…