Update README.md

README.md

@@ -1,8 +1,8 @@
 ---
-title: The Seamstresses
-emoji: ✂️
-colorFrom: blue
-colorTo: green
+title: Metabolic Phoenix Framework
+emoji: ⚡
+colorFrom: green
+colorTo: yellow
 sdk: gradio
 sdk_version: 6.11.0
 app_file: app.py
@@ -10,60 +10,80 @@ pinned: true
 license: other
 ---
 
-# ✂️ The Seamstresses
+# ⚡ The Metabolic Phoenix Framework
 
-**Intelligent Document Deduplication & Best-Content Selection**
+**Forensic Document Analysis · Legal Document Generation · Neural Voice Readback**
 
-Built by **Dwayne Anthony Brian Galloway** · Patent Pending
+Built by **Dwayne Anthony Brian Galloway** · Injectiv8 · Ms. Nala · Patent Pending
 
 ---
 
 ## What It Does
 
-The Seamstresses takes multiple documents covering similar or overlapping content — emails, letters, statements, evidence bundles — and produces a single, clean, professionally structured master document containing only the highest-quality, non-redundant information.
+The Metabolic Phoenix Framework analyses documents containing evidence of institutional breaches and produces a full forensic case record across four cycles — fact extraction, compliance analysis, pattern detection, and litigation-ready case output. It then generates five court and regulatory documents from that analysis, ready to send.
 
-No content is added, inferred, or associated beyond what is present in the source files.
+---
+
+## The Four Cycles
+
+**Cycle 1 — Fact Ledger (Introbot)**
+Extracts every confirmed fact, admission, denial, contradiction, and omission from the documents. States what is on the record without hedging.
+
+**Cycle 2 — Compliance Analysis (RHB 2.0)**
+Maps every institution to every statutory duty they held. Confirms each breach by statute, section, and date. Traces the full causal chain from first breach to final harm.
+
+**Cycle 3 — Ouroboros Loops (Phoenix Badger)**
+Runs four analytical rings: evasion tactics and bad faith (Ring A), past unlawful decisions (Ring B), patterns of institutional misconduct (Ring C), internal contradictions (Ring D). Closes with an escalation map.
+
+**Cycle 4 — Full Case Output (Layoutbot)**
+Produces a complete litigation-ready case summary: executive summary, confirmed breaches, top five arguments, required next steps, and escalation map.
+
+---
+
+## Document Generation Engine
+
+After running the analysis, generate five court and regulatory documents at the click of a button:
+
+1. **Letter Before Action** — Pre-Action Protocol, 14-day deadline, litigation follows without further notice
+2. **Formal Regulatory Complaint** — for FOS, FCA, ICO, or Parliamentary Ombudsman
+3. **Witness Statement** — first person, court-ready, King's Bench Division heading, Statement of Truth
+4. **Skeleton Argument** — structured legal argument with authorities, issues, and relief sought
+5. **Regulatory Referral** — FCA, SRA, or ICO referral with 28-day deadline and escalation threat
 
 ---
 
 ## Features
 
-- **Intake & Extraction** — PDF, DOCX, and TXT supported
-- **Similarity Engine** — 85%+ threshold paragraph-level deduplication
-- **Best-Content Selection** — retains the clearest, fullest version of repeated content
-- **Section Classification** — auto-organises output into logical headers
-- **Risk Language Detection** — identifies denial traps, admission by silence, procedural hedging, burden-flipping, and smoking guns
-- **SHA256 Integrity Seal** — cryptographic proof of every output
-- **Audit Trail** — full provenance metadata tracking source document for every selected paragraph
-- **Export** — master document as `.txt` and full audit as `.json`
-- **Text-to-Speech** — read any output aloud
+- **Multiple file formats** — PDF, images (OCR), audio (transcription), video, text
+- **Multiple documents** — upload as many as needed, all analysed together
+- **Neural Voice Readback** — Microsoft Sonia (en-GB) neural voice, full length, no time limit
+- **Speed and Pitch Controls** — adjust reading speed and voice pitch
+- **Read Document Only** — read any uploaded document aloud without running analysis
+- **Save and Download** — analysis and all five documents as downloadable text files
 
 ---
 
 ## How To Use
 
-1. Go to the **Intake** tab and upload your documents
-2. Click **Ingest Documents**
-3. Go to the **Run Seamstresses** tab and click **Run Seamstresses**
-4. Review the **Master Document**, **Risk Language**, and **Audit Trail** tabs
-5. Export your files from the **Export** tab
+1. Upload your documents using the file upload box at the top
+2. Click **Analyse Documents** to run all four cycles
+3. Read or listen to each cycle's output
+4. Scroll to the Document Generation Engine
+5. Click **Generate ALL Documents** or generate each one individually
+6. Download everything using the save buttons
 
 ---
 
-## Use Cases
+## File Types Supported
 
-- Legal evidence consolidation
-- Deduplicating correspondence bundles before court submission
-- Merging multiple versions of the same document
-- Identifying risk language and admissions in institutional responses
-- Any situation where multiple near-identical documents need to become one clean file
+PDF · DOCX · TXT · JPG · PNG · MP3 · WAV · MP4 · MOV · and more
 
 ---
 
-## Part of the HoneyBadger Ecosystem
+## Part of the WhiteHat Ecosystem
 
-The Seamstresses is the evidence-cleaning layer of a broader cross-regulatory litigation infrastructure.
+The Metabolic Phoenix Framework is the primary forensic analysis and document generation layer of a broader cross-regulatory litigation infrastructure built by Dwayne Anthony Brian Galloway.
 
 ---
 
-*© 2025 Dwayne Anthony Brian Galloway · All Rights Reserved · Patent Pending*
+*© 2026 Dwayne Anthony Brian Galloway · Injectiv8 · All Rights Reserved · Patent Pending*