Discover
Start from focused recipe, medicine-making, monograph and method collections. Only links matching a configured profile enter the queue.
Research intelligenceBroad research input.
Broad research input.
One editorial standard.
The system separates discovery, research, drafting and publication. A successful crawl creates internal research lineage — never an automatically copied public page.
The complete pipeline
Six stages, six different jobs.
Keeping these layers separate prevents archive noise, duplicates and unlabelled health claims from leaking into the public journal.
Classify
Collections, articles, gated pages, promotions and unrelated pages are scored differently. Navigation pages discover links but do not become editorial notes.
Fingerprint
Canonical, title, exact-content, recipe and similarity signatures are calculated before drafting.
Cluster
Exact and near-identical pages attach to one internal research topic, preserving correction and audit history without exposing source-site promotion on public notes.
Structure
Ingredients, methods, botanical mentions, safety language and preparation metadata are normalized into the HHT data model.
Publish
Public notes display the preparation, evidence label, safety boundary, herb links and independent citations where available.
What each command createsCrawl is research.
Crawl is research.
Publish is editorial.
The distinction keeps the private research lineage intact while public pages remain clean, original and useful.
Source-profile architecture
Six independent profile types
Profiles remain internal configuration. The public journal presents HHT content rather than advertising the research intake sites.
- Education archivePaginated remedies, recipes, plant profiles and beginner material.
- Clinical editorial archiveRecipe and medicine-making collections with structured article paths.
- Practitioner archiveKitchen-apothecary hubs leading to dated practical articles.
- Method archivePreparation hubs, technique pages and partial-content detection.
- School archiveRoot-level recipe and medicine-making articles with course and promotional exclusions.
- Commercial editorial archiveBroad recipe collections requiring both link-context and on-page herbal signals.
!
Non-negotiable editorial rule
Read the evidence framework ↗Research input is not proof by itself.
Preparation archives can establish technique and usage context. Clinical or regulatory claims require a separate authority layer and an explicit evidence label.