Developer Documentation

These are the developer resources for subramanya.ai, the personal site of Subramanya N. This section is the human-readable counterpart to the machine-readable /openapi.json, /.well-known/api-catalog, and /llms.txt endpoints. It describes how the site is built, what data it depends on, how agents can consume it programmatically, and how to contribute.

Pages

  • Architecture - system overview, build flow, rendering model, and where to change things.
  • Content Model - front-matter and data contracts for posts, books, data files, and top-level pages.
  • Development - local setup, validation commands, content update workflow, and Python maintenance scripts.
  • Deploy Checklist - pre-deploy verification: production build, content/link validation, API-vs-spec checks, and post-deploy live checks.
  • API Versioning & Deprecation Policy - URL path versioning, additive-only guarantees, 6-month sunset timeline, and how deprecations are announced.

Agent-Readable Endpoints

For automated consumption, the site exposes several machine-readable endpoints. These are also advertised via RFC 8288 <link> elements on every page and via an RFC 9727 api-catalog Linkset.

EndpointPurposeContent Type
/openapi.jsonOpenAPI 3.1 spec of the read-only content APIapplication/json
/api/v1/site.jsonSite metadata + endpoint directory (versioned, typed JSON)application/json
/api/v1/posts.jsonAll blog posts (versioned, typed JSON)application/json
/api/v1/books.jsonAll book notes (versioned, typed JSON)application/json
/api/v1/tags.jsonAll topic tags with archive URLs (versioned, typed JSON)application/json
/.well-known/api-catalogRFC 9727 catalog of all machine-readable endpointsapplication/linkset+json
/llms.txtShort site summary for LLM ingestiontext/plain
/llms-full.txtFull site content for LLM ingestiontext/plain
/search.jsonSearch index (posts and books)application/json
/feed.xmlBlog RSS 2.0 feedapplication/rss+xml
/sitemap.xmlSitemaps 0.9application/xml
/sitemapindex.xmlSitemap indexapplication/xml
/robots.txtCrawler policy with AI content signalstext/plain

All endpoints are read-only, served over anonymous HTTPS GET - no API key, no sign-up, no rate-limit registration. Every operation in /openapi.json has a unique operationId, a description, and typed response schemas, so it can be loaded directly into LLM function-calling or tool-use frameworks. Nonexistent paths return a real HTTP 404 whose body links to /llms.txt, /sitemap.xml, and /search.json for recovery.

RFC References

  • RFC 8288 - Web Linking (the Link header and <link> element).
  • RFC 9264 - Linkset format (application/linkset+json).
  • RFC 9727 - API catalog well-known URI.
  • RFC 8615 - Well-known URIs.

Source

The complete source of this site — including the agent configs (AGENTS.md, plugin.json, skills/) — lives at github.com/subramanya1997/subramanya1997.github.io.