MeddAI - Medical Content Creator AI Web Application
Medd AI Web App
A web app that drafts compliant medical content tuned to brand voice, with built-in approval and publishing workflows.
MeddAI writes medical content that carries its sources, for doctors, for students, and for readers with no training at all. Behind it sits a Python and FastAPI service that runs several AI agents through LangGraph. Google Gemini and OpenAI GPT-4o do the writing and the checking; where a request arrives with no references of its own, Perplexity AI goes and finds them.
You can hand it a link, a Word file, a PDF, or a YouTube video. It pulls the sources out as it reads, checks the medical claims against them, and returns an article set in AMA style. There is a REST API as well, so a request can come from a person or from another system.
- AI-powered medical content generation
- Citation-ready healthcare articles
- FastAPI backend
- LangGraph AI orchestration
- Google Gemini integration
- OpenAI GPT-4o validation
- AMA formatting
- Medical accuracy verification
- Multi-source ingestion (URL, PDF, DOCX, video)
- REST API healthcare
Solving real problems with smart engineering
Validating Medical Accuracy
Problem
Risk of generating inaccurate or non-compliant content that could mislead healthcare audiences.
Solution
Implemented GPT-4o for post-generation validation - checking factual accuracy, AMA citation format, and readability. Invalid drafts are automatically flagged and returned with issue logs.
Ingesting Multiple Content Types
Problem
Required support for diverse sources like PDFs, DOCX, web pages, and YouTube videos in a single pipeline.
Solution
Integrated Firecrawl for web pages, PyPDF2 / python-docx for documents, and RapidAPI for YouTube transcription - seamless ingestion and parsing across formats.
Dynamic Audience Adaptation
Problem
One-size-fits-all tone fails across user groups - doctors, students, patients, and the public all need different language.
Solution
Language generation dynamically adjusts tone and jargon to fit doctors, students, patients, or the general public based on the request context.
Citation-Enforced Content
Problem
Generated content lacked scholarly reference formatting - making it unusable for evidence-based publication.
Solution
Google Gemini is instructed with strict prompting to embed in-text citations and meet academic AMA formatting standards.
Inside Medd AI Web App
Built with
Frontend
- Any client (React.js, Postman, etc.) - sends requests and renders article output
- REST endpoints designed for easy integration into existing UIs
Backend
- FastAPI - endpoints for article generation, extraction, and health checks
- LangGraph - orchestrates the multi-step workflow (input → research → generation → validation)
- Python - core business logic and integrations
AI Models & External APIs
- Google Gemini - generates draft content with strict citation prompting
- OpenAI GPT-4o - validates factual accuracy and formatting compliance
- Perplexity AI - literature search fallback when no source is provided
- Firecrawl, PyPDF2 / python-docx, RapidAPI - multi-format ingestion
Recommendation Engine & Security
- Auto-routes requests to the right workflow based on input type (URL, PDF, DOCX, video)
- Word-count accuracy within ±10% via Gemini constraints
- Asynchronous validation and rate-limiting for abuse protection
- API keys behind secure access layers; HIPAA-conscious practices
What actually changed
Have a similar project in mind?
Let's talk about how AI can transform your service business.
Discuss Your Idea Today
