Xandrov StudioAtlas of shipped systems
← The territory/ V. Brand & Web/ № 18 — Research
Dossier № 18 / 19 — Brand & Web — Live production site · Pipeline

Xandrov Research

Book-quality reading ergonomics from plain Markdown.

File
№ 18 / 19
Territory
Brand & Web
Classification
Live production site · Pipeline
Status
Live in production
§ 01 Context

Xandrov Labs publishes long-form strategy content — two companion volumes intended for serious, extended reading. These documents demand book-like ergonomics: readable typography, clear navigation, dark mode, progress tracking. But the author should not be forced to think about design or deployment; the source of truth must remain plain Markdown. The system exists to close this gap, delivering professional reading experience without authoring friction or manual deployment steps.

§ 02 The problem

Long-form documents deserve book-quality reading ergonomics — sticky table of contents, reading progress indicators, light and dark modes, styled tables and callouts — yet generating such output by hand is error-prone and slow. The challenge: keep the source in plain Markdown — the author's native format — while shipping polished, feature-complete HTML. This requires a build system that handles both the ergonomic requirements and end-to-end automation, without placing extra burden on the writer.

§ 03 The approach

A Node build tool reads Markdown files from docs/ and renders them into a complete static site. The tool generates not only content pages but also index, robots.txt, response headers, and 404 error handlers. Every push to main triggers CI to rebuild and deploy automatically to production.

The rendering pipeline produces sticky-scrolling table of contents, reading-progress indicators, light-and-dark theme support, and styled semantic elements for tables, callouts, and footnotes. All features flow from Markdown source — there is no hand-coded HTML in the output. The build process is deterministic and repeatable, ensuring the rendered site always reflects the current state of source documents.

Book-quality reading ergonomics from plain Markdown.

§ 04 Decisions that mattered
D1
Single source of truth
Only docs/*.md is hand-edited by authors. All output — site pages, index, headers, error pages — is generated by the build tool. This ensures consistency and eliminates manual drift.
D2
Generated output is immutable
The site directory is never touched by hand. Every file is produced by the build tool, making deployments repeatable and the build process the single authoritative step.
D3
Noindex headers for deliberate unlisting
Content is served with noindex headers — not secrecy, but deliberate strategy to keep the site out of search results.
D4
Automated deployment on every push
CI/CD builds and deploys the site automatically whenever main is updated. Authors commit Markdown; the system handles the rest.
§ 05 Where it stands

The system is live in production at rs.xandrov.com, deploying automatically on every push to main via CI/CD. Authors shift focus entirely to content — writing and refining Markdown — while the pipeline ensures readers consistently receive polished, book-quality reading ergonomics without manual intervention. The entire workflow from commit to deployment is hands-off.

Node build toolMarkdown → HTMLCI/CDStatic host

LIVE — rs.xandrov.com ↗

§ 06 — Sanitization record

The description focuses on the technical architecture and deployment pipeline. Specific content, internal performance metrics, and proprietary build configuration details are not included. The system's strategy-content purpose is noted but the volumes themselves remain private to the lab.

Have a system like this worth building?

Fixed scope, quoted after one working session, first output inside two weeks. Read by a person within one working day.