Orris

What you do that gpt dont

Here's what sets me apart from ChatGPT (GPT-4/4o):

What I do that GPT doesn't (or does differently)

1. Medical Textbook Library

I have direct access to a curated library of medical textbooks. When you ask a clinical question, I search actual textbook content and cite specific pages - not just my training data. GPT answers from memory; I pull from indexed sources.

2. PubMed Integration

I can search PubMed in real time, retrieve abstracts, download full-text PDFs, and cite peer-reviewed studies with PMIDs. GPT can't access live PubMed.

3. Live Code Execution (Sandbox)

I run code in a persistent Linux sandbox - Python, shell scripts, file generation, data processing. The results persist between turns. GPT also has a code interpreter, but my sandbox is a full Linux environment (not restricted).

4. File Creation & Download

I can generate real downloadable files - .pptx, .docx, .xlsx, .pdf - and give you a direct download link. GPT can generate files too, but my skill-based system uses structured workflows to produce polished, multi-slide or multi-section documents.

5. Specialist Skills

I have domain-specific skill modules:
  • Indian Taxation (CA Bhanwar Borana's materials - Direct Tax, Transfer Pricing)
  • GST / Indirect Tax (Navkar Institute - CA Final, ACCA prep)
  • Orthopaedic Surgery (hip surgery videos + clinical guidelines)
  • Medical Library (full textbook search with images)
GPT has broad knowledge; I have deep, structured, citable knowledge in these specific domains.

6. Mandatory Clarification Before Acting

Before I tackle ambiguous or broad questions, I ask targeted clarifying questions to avoid wasted effort. This saves time on complex tasks.

7. Parallel Tool Execution

I run multiple searches, file reads, and computations simultaneously in a single turn - not one at a time. This makes complex research faster.

8. Structured Task Management

For long multi-step tasks, I use a visible to-do list to track progress, which keeps complex work organized and transparent.
In short: I'm optimized for depth in specific domains (medicine, Indian tax/GST, document generation) with real file I/O, live literature search, and a persistent coding environment. GPT is broader and more general-purpose. If your work touches those domains, I'll likely outperform it significantly.

Retry

Here's what I do that GPT doesn't:

What I do that GPT doesn't

1. Real Medical Textbook Search

I search an actual indexed library of medical textbooks and cite specific pages. GPT recalls from training memory - I pull from the source.

2. Live PubMed Access

I search PubMed right now, fetch abstracts, and download full-text PDFs with real PMIDs. GPT has a knowledge cutoff and no live literature access.

3. Persistent Linux Sandbox

My code runs in a real Linux environment that persists across turns. Files I create stay there. GPT's interpreter is more restricted and resets.

4. Real Downloadable Files

I generate .pptx, .docx, .xlsx, .pdf files you can actually download via a direct link - not just view in a chat window.

5. Domain Skill Modules

I have structured, deep knowledge modules for:
  • Indian Direct Tax (CA Final - Bhanwar Borana materials)
  • GST / Indirect Tax (Navkar Institute - CA Final, ACCA)
  • Medical library with textbook images
  • Orthopaedic hip surgery (surgical videos + guidelines)

6. Parallel Tool Execution

I fire multiple searches, reads, and computations at the same time in one turn. Faster for complex research.

7. Mandatory Clarification

On broad or ambiguous tasks, I ask before I act - so I don't waste your time producing something you didn't want.
Bottom line: GPT is wide. I'm deep - especially in medicine, Indian taxation, and document generation, with real file I/O and live literature.
This is a shared conversation. Sign in to Orris to start your own chat.