OWEN
OncWonk Evidence Nexus
An advanced AI-powered literature synthesis and hybrid RAG research engine evaluating 130,000+ curated peer-reviewed integrative oncology studies.
The Integrative Oncology Evidence Gap
Clinical researchers and oncologists face three fundamental bottlenecks when evaluating complementary therapies:
📈 Information Overload
Over 130,000+ published studies on natural compounds, botanicals, and molecular pathways exist—scattered across fragmented databases.
⚠️ Hallucination Risk
Generic consumer AI models fabricate non-existent trials or conflate in vitro cell studies with human clinical results.
⏱️ Clinical Time Crunch
Clinicians lack the hours needed to manually review dozens of papers, compare trial dosages, and evaluate drug-herb interactions.
Why OWEN is Different
Engineered from the ground up for strict clinical precision, verifiable citations, and high throughput:
Pre-indexed database focused exclusively on botanicals, nutraceuticals, vitamins, minerals, and integrative therapeutics.
🛡️ Null-Intersection Protocol
If a compound and tumor type have zero empirical data, OWEN explicitly alerts you to the data deficit rather than fabricating synthetic claims.
⚡ Google Gemini Context Caching
Pre-caches entire paper cohorts in memory for instant, cost-effective multi-turn clinical synthesis.
Ultra-Fast Hybrid Search Engine
Combines Boolean exactness with 1536-dimensional conceptual AI understanding.
1. Keyword (Exact)
Full Boolean algebra with uppercase AND, OR, NOT, quotes, and nested ().
2. Semantic (Vector)
Discovers conceptual mechanisms and molecular pathways even when exact keywords differ.
3. Hybrid (RRF)
Executes both engines and fuses ranks using Reciprocal Rank Fusion for optimal precision.
Automated MeSH Synonym Expansion
Bridges the gap between trade names, botanical binomials, and clinical abbreviations.
Automatic Concept Recognition
OWEN recognizes hundreds of clinical abbreviations. Typing AHCC automatically searches for Active Hexose Correlated Compound.
Expands To: Curcuma longa, diferuloylmethane, turmeric extract
Interactive MeSH Suggestion Badges
Suggested MeSH tags appear above search results. Clicking any badge injects that term directly into your active search with 1 click.
Clarification Cards & Cancer Focus
Guiding research cohorts with intelligent multi-modal filtering.
Dynamic Clarification Cards
When queries are broad, OWEN surfaces interactive suggestion pills right above the results.
[ER/PR+] [HER2+] [Triple-Negative]
21-Category Cancer Focus Selector
Instant dropdown selector in the top bar covering Glioblastoma, Colorectal, Prostate, Melanoma, Pancreatic, Ovarian, Leukemia, and more.
Precision Cohort Scoping
Define the exact empirical evidence boundary before synthesizing or querying:
🎯 Dynamic Context & Scaling
Select specific studies or an entire cancer category with thousands of papers via the Cancer Focus selector. OWEN dynamically scales reasoning and token ceilings to evaluate the full cohort.
🛡️ Defined Corpus Requirement
To maximize clinical precision and guard against hallucinations, un-filtered queries across the open 130,000+ database are disabled.
Clinicians define their active corpus via search selection or the Cancer Focus dropdown.
The AI Synthesis Suite
Generates publication-grade literature syntheses in seconds using 6 curated blueprints:
✨ Patient Explainer
Translates complex trial data into clear, accessible patient summaries.
📖 Literature Review
Reviews grouped by molecular mechanisms, study tiers, and clinical outcomes.
📊 Key Findings
Bulleted takeaways, evaluated trial dosages, and high-impact clinical stats.
🔬 Methodology
Side-by-side analysis of sample sizes, human vs. animal models, and designs.
⚖️ Comparative
Compares and contrasts 2+ substances, mechanisms, or protocols.
📋 Protocol Review
Evaluates an uploaded protocol PDF or regimen against the evidence base.
Evidence Hierarchy & Molecular Pathways
Transparent mechanistic modeling grounded in established oncology evidence hierarchies.
Strict Evidence Stratification
All syntheses systematically prioritize evidence by clinical weight:
- Human Clinical Trials & Reviews
- Animal In Vivo Models
- In Vitro Cell Line Studies
- Molecular Pathway Modeling
Interactive Mermaid Pathways
Biochemical signaling cascades automatically render as interactive visual diagrams.
Includes a Full-Screen Viewer for inspecting intricate multi-step molecular pathways.
OWEN Chat & Protocol Uploads
Multi-turn clinical Q&A with deep drill-down actions directly beneath syntheses.
💬 Multi-Turn Grounded Chat
Interactive conversational RAG grounded strictly in your peer-reviewed library with clickable blue citation pills.
“Evaluate this patient’s 4-botanical protocol against our glioblastoma cohort for safety and synergy.”
📎 PDF Attachments & [🔬 Dig Deeper]
Upload lab/supplement PDFs directly into threads.
Click [🔬 Dig Deeper] beneath answers for immediate high-budget drill-downs into dosages and clinical trial tiers.
Deep Research & Prompt Extender
Automating iterative evidence compilation and query engineering.
✨ Prompt Extender (Magic Wand)
Type a basic idea (e.g. “green tea prostate”) and click the Wand icon.
OWEN automatically expands it into rich clinical prose with MeSH terms and mechanistic targets.
🔬 Autonomous Deep Research
Launches a 4-stage, multi-stream retrieval workflow executing iterative passes across your library to compile an exhaustive meta-analysis in ~30–45s.
1-Click Export & Saved Consultations
Seamless integration into clinical records and research publications.
📋 Copy for Export
Converts internal app links into publication-ready inline citations for Word or Docs:
🔖 Consultations & Folders
Save research threads, organize consultations into folders by patient or cancer type, and reload past sessions with selected papers intact.