Create README.md

README.md

@@ -0,0 +1,49 @@
+---
+license: mit
+task_categories:
+- text-classification
+- question-answering
+- tabular-classification
+tags:
+- biology
+- chemistry
+- medical
+- botany
+- json
+- api
+pretty_name: Cleaned USDA Phytochemical Database (Sample)
+size_categories:
+- n<1K
+---
+
+
+# USDA Phytochemical Database (Cleaned JSON)
+
+This repository contains a **free 400-record sample** of the highly unstructured USDA Dr. Duke's Phytochemical and Ethnobotanical Databases, transformed into a clean, relational, and machine-readable JSON format.
+
+## The Problem
+Historically, accessing and structuring pharmacological, ethnobotanical, and toxicological plant data required hundreds of hours of manual CSV scraping, data cleansing, and mapping. 
+
+## The Solution
+I have completely cleaned, normalized, and mapped the entire dataset. 
+
+If you are a Data Scientist, building Pharma R&D pipelines, or developing AI Compliance Agents for Nutraceuticals, you do not need to waste time cleaning data. 
+
+**The full database contains 104,388 exact phytochemical records (2,315 plant species, 24,771 chemicals).**
+
+### ⚡ Get Full Access & High-Speed API
+Instead of scraping, you can get instant access to the complete 104k+ dataset via a Sub-Millisecond REST-API and MCP (Model Context Protocol) Server for LLMs.
+
+🔗 **[Get Instant API Key & Full Dataset Access (€99/mo)](https://buy.stripe.com/4gM6oG74HcAXcdI2HKebu00)**
+
+*Zero maintenance. Zero friction. Ready for production pipelines.*
+
+## Sample Structure
+The included `usda_phytochemical_sample_400.json` demonstrates the clean entity mapping:
+```json
+{
+  "chemical": "patentiflorin a",
+  "plant_species": "justicia gendarussa",
+  "application": "hiv inhibition",
+  "dosage": "high"
+}