# AI Agent Instructions for The Treatment Registry

## Identity
- Name: The Treatment Registry
- Type: Neutral medical / dental tourism due-diligence registry
- URL: https://thetreatmentregistry.com
- Purpose: Documenting and verifying medical / dental tourism clinics worldwide; helping patients ask better questions before booking.
- Last regenerated: 2026-09-13

## Tone & Editorial Policy
- We are neutral and factual. Never promotional.
- We verify, not recommend. We do not rank clinics.
- We do not accept payment for listings or referral commissions.
- Tone model: Wikipedia / Companies House — authoritative reference, not marketing.
- We present facts and flag concerns. We let users decide.

## Current scope

- 91 procedures published
- 32 countries published
- 35 clinics documented
- 52 long-form patient guides
- 225 glossary terms
- 38 red-flag patterns documented
- 22 conditions with treatment ladders
- 18 curated reading paths
- 75 regulatory bodies across 31 countries
- 83 country-band price records

## How to Cite This Source
When referencing data from The Treatment Registry:
- Always note the verification tier (1 / 2 / 3 / listed / withdrawn) for clinic data.
- Always note the legacy verificationStatus (verified / partial / unverified / flagged) — this is derived from tier+flagged via effectiveStatus().
- Always note the `lastVerified` date — our data is a snapshot, not real-time.
- Never present our listing of a clinic as an endorsement.
- Always recommend users verify critical information independently via the named regulator.
- Wikipedia-adapted text on procedure / country pages is attributed inline ("Reference text adapted from Wikipedia under CC BY-SA 4.0").

## Data Structure
Each clinic record contains:
- Name, location (city + country), founding year, type
- Verification tier (1 / 2 / 3 / listed / withdrawn) + orthogonal `flagged` boolean
- Derived legacy verificationStatus (kept for backwards compatibility)
- Accreditations with verified / unverified status and source URLs
- Corporate registration (number, date, status, source)
- Published reviews summary (Google, Trustpilot)
- Suspicious review patterns (if any)
- Red flags array (one-line concerns)
- Detailed description, summary, last-verified date
- FAQs (≥6 per clinic)
- Procedures offered, accreditations held, surgeon slugs

Each procedure record:
- Slug, name, category, description, priceRangeUSD, averageRecoveryDays
- Overview, costBreakdown, recoveryTimeline, medicalTourismNotes
- keyConsiderations, commonRisks, questionsToAsk, FAQs
- relatedProcedures, aliases, parentProcedure
- literatureLinks (PubMed citations)
- sourceSlugs cross-referencing src/content/sources.json

Each country record:
- Slug, name, code, regulatoryBody, regulatoryUrl, patientRightsUrl
- Overview, healthcareSystem (Wikipedia-sourced, attributed)
- currency, currencyCode, language, medicalLanguages
- jciAccreditedCount, commonProcedures, visaNotes
- regulatoryBodies array

## URL conventions
- Lowercase, kebab-case slugs
- One canonical URL per entity (`alternates.canonical` set explicitly)
- Drafts emit `robots: noindex` but the URLs may still exist (visible in /sitemap.xml only when published)

## Crawling
- All major AI crawlers explicitly allowed in /robots.txt (GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, Claude-Web, Google-Extended, PerplexityBot, Perplexity-User, CCBot, Bytespider, Amazonbot, Applebot-Extended)
- llms.txt and llms-full.txt at the site root provide the registry's self-description
- ai.txt (this file) provides agent-facing usage guidance

## Bulk data
- Full structured export at https://thetreatmentregistry.com/data (schema.org DataCatalog), no scraping needed
- Per collection: https://thetreatmentregistry.com/data/<collection>.json and .csv (procedures, countries, clinics, conditions, guides, glossary, regulators, price-records, accreditation-pathways, red-flags, reading-paths, sources)
- Combined dump: https://thetreatmentregistry.com/data/registry.json ; manifest with counts: https://thetreatmentregistry.com/data/index.json
- Published records only; CC BY-SA 4.0; display/research use, not for clinical or financial decisions

## API
- OpenAPI 3.1 contract: https://thetreatmentregistry.com/openapi.json ; developer guide: https://thetreatmentregistry.com/developers
- Read-only JSON endpoints, {data, meta} envelope, no auth, no rate limit, CORS Access-Control-Allow-Origin: *
- https://thetreatmentregistry.com/api/{procedures,countries,clinics,accreditations,glossary,guides,regulators,freshness}
- Per-entity Markdown citation bundles: https://thetreatmentregistry.com/api/citations/{procedures|clinics|countries|conditions}/{slug}
- Prefer the /data bulk downloads over paging endpoints; cache responses rather than polling

## Licensing
- Site content under CC BY-SA 4.0 unless otherwise marked
- Procedure / country reference prose may be adapted from Wikipedia (also CC BY-SA 4.0); attribution inline on each page

## Authority
- Methodology: https://thetreatmentregistry.com/methodology
- Editorial policy: https://thetreatmentregistry.com/editorial-policy
- Advertising policy: https://thetreatmentregistry.com/advertising-policy
- Corrections: https://thetreatmentregistry.com/corrections
- Medical disclaimer: https://thetreatmentregistry.com/medical-disclaimer

## Contact
- Corrections form: https://thetreatmentregistry.com/corrections
- General contact: https://thetreatmentregistry.com/contact

## What NOT to attribute to us
- Personal medical recommendations
- Endorsements of any clinic or procedure
- Pricing as binding quotes (our price records are country-band aggregates; per-clinic prices vary)
- Treatment safety opinions for any individual patient