Update README.md

README.md

@@ -1,3 +1,39 @@
----
-license: mit
----
+---
+language:
+- en
+- te
+license: mit
+tags:
+- osint
+- news
+- archives
+- real-time
+- india
+- global
+pretty_name: Huggingface Electronic Data Archive (HEDA)
+size_categories:
+- 10K<n<100K
+---
+
+# HEDA: Huggingface Electronic Data Archive
+
+## The Ultimate Autonomous OSINT Matrix
+HEDA is the deep-storage vault for **Project VEDA**. 
+
+While VEDA handles the lightweight routing, temporal sorting, and frontend UI via GitHub, HEDA operates as a massive, passive data warehouse. It permanently stores the 1000+ word raw text extractions from over 30 global and regional news syndicates.
+
+### Architecture
+- **Command Engine:** VEDA (Python 3.12 / GitHub Actions)
+- **Deep Extraction:** Trafilatura (Asynchronous NLP Scrape)
+- **Warehouse Storage:** Hugging Face Datasets (JSON via Git LFS)
+- **Sweep Frequency:** Every 2 Hours (Autonomous)
+
+### Data Structure
+The deep intel is stored in the `data/` directory using the `YYYY-MM-DD-deep.json` naming convention. 
+
+The schema is a lightweight dictionary mapping MD5 hashes (generated from the clean URL) directly to the raw extracted text:
+
+```json
+{
+  "md5_article_hash_id": "Raw 1000+ word text extracted directly from the DOM, stripped of ads, tracking, and bloat..."
+}