New .claude/rules guidance so your project's AI agent knows about private (gated) collections, subdomain hubs, and that locking access (Cloudflare Access) is a separate manual step. Plus docs for setting it up.
Documentation: how to deploy the Trust Center (or any collection) to a subdomain, where the root renders the collection's own index template.
Bugfix: a collection deployed to a subdomain now renders its root page with the collection's own index template (e.g. the data-driven Trust Center hub) instead of the generic site index.
Data file values can now carry per-language variants. A value like { en: "Cloud infrastructure", de: "Cloud-Infrastruktur" } resolves to the current page language, so the bundled Trust Center is bilingual out of the box. Plain single-language data is unchanged.
New `private = true` collection flag builds a collection's hub and pages but keeps them out of every public discovery surface (sitemap, llms.txt, search index, feeds, homepage) — for content placed behind Cloudflare Access or HTTP auth.
seite now collects anonymous, opt-out usage telemetry (which commands run, on what platform) to guide development. No paths, content, or personal data are sent — disable any time with `seite telemetry off`, SEITE_TELEMETRY=0, or DO_NOT_TRACK=1.
Bug-fix release: the MCP search tool no longer crashes on non-ASCII content, frontmatter parsing handles `---` inside values, self-update verifies checksums in-process, and GitHub Pages re-deploys are idempotent.
Maintenance release: dependency updates, a transitive security patch, and CI fixes. No behavior changes.
Auto-generate subdomain deploy steps in CI workflows and stop pinning the seite version so projects always use the latest release.
Fix AVIF image processing errors and add i18n support for shortcode callout labels.
Plausible Analytics: support new unique snippet format (Oct 2025+) with async loading and init script.
Security fix: heading XSS, i18n language context bug, Plausible warning spam, image attribute extraction
Incremental builds: dev server skips rebuilding unchanged content for faster iteration
Bug fixes: PageSpeed URL encoding, Atom feed RFC 4287 author compliance, tag link URL prefix, CLAUDE.md pagination example
Fix serve watcher rebuild loop on Linux with strictatime/relatime filesystems
Atom 1.0 feeds, redirect aliases via frontmatter, upgrade path for custom templates
Previous/next article navigation for blog posts
PageSpeed audit command, CSS/JS minification on by default
Path-scoped .claude/rules/ context files for leaner CLAUDE.md
Major UX improvements: build progress spinners, Vite-style dev server, shell completions, smart error hints, and more.
Collection index pages with sidebar nav, redirect_to, and docs-index template
Fix seite upgrade not updating project version when no file changes needed
Per-collection subdomain deploys
Further build pipeline optimizations: eliminate redundant file walks and per-page allocations.
Per-subdomain base URL override
Parallelize the build pipeline with rayon and consolidate HTML post-processing for faster builds.
Subdomain support in MCP, scaffolding, and upgrade system
Fix Windows build by switching KaTeX from quick-js to duktape backend. The quick-js JavaScript engine does not support Windows targets, causing release builds to fail.
4 new themes, 2026 design refresh, math/LaTeX rendering, AVIF image format, and SEO improvements.
Brand identity builder skill: logo, color palette, and favicon creation via /brand-identity.
Plausible script extensions, conversion event tracking, and PNG image optimization.
Image quality settings now apply to JPEG and WebP output: lossy WebP encoding via libwebp.
SEO meta tag fixes across all themes and new public/ directory for root-level files.
Initial public release of seite, the AI-native static site generator built in Rust.