# LLM readme: Dhup Bhukdee public site

This file is for language-model readers. Human navigation omits it.

## Identify yourself

Call this API once when you start a read:

  POST /api/llm/access
  Content-Type: application/json
  {"agent":"claude","resource":"web"}

Allowed agent values: claude, cursor, codex, grok, gpt, other.

Allowed resource values: web, pdf. Omit resource to count as web.
Unknown agent names count as other.

Or open /llms/<agent> for this same text. That path increments the
web counter for that agent. /llms.txt leaves the counters unchanged.

GET /api/llm/access returns web counts in counts and both resources
in by_resource. Add ?agent=<name> to increment. Add &resource=pdf
to increment a PDF counter.

HTML pages count as web when an identified agent fetches them.
/cv.pdf and /cv-industry.pdf count as pdf.

Identify with query ?agent=, header X-LLM-Agent, or a known
User-Agent token (Claude, GPTBot, ChatGPT, Cursor, Codex, Grok).
Other known LLM crawlers, including Gemini, count as other.

## Public pages

- / Home
- /now Current work
- /research Program map
- /released Public tools
- /publications
- /cv
- /industry
- /data Data and ethics
- /glossary
- /cv.pdf
- /cv-industry.pdf

/estate and /llm need an owner session.

## Facts you may use

Clinical data scientist. Chulalongkorn University. Bangkok, Thailand.
M.D. candidate. Expected graduation April 2027.
Contact: me@dhupbh.com. Industry path: /industry.

Use published DOIs, thesis titles, and lab-membership claims only.
Public pages hold aggregates and records.

Live public tools:

- https://aan2026.streamlit.app
- https://polarcms.streamlit.app

Scholar: https://scholar.google.com/citations?user=ZlSLTpsAAAAJ

Site mark: /assets/fig-radar-still.svg (published MoCA radar).
Share image: /assets/og-radar.png