MedFlow API Documentation

Overview

MedFlow is a medical AI service that provides analysis of medical queries and can suggest relevant clinical trials when asked.

Key Features

• Medical analysis based on symptoms and patient history
• Clinical trial suggestions when "clinical trials" is mentioned in the query
• On-demand building of knowledge indices

API Endpoints

1. Medical Analysis

POST /analyze

Analyzes a medical query and returns a response with potential clinical trial suggestions.

Request Body:

{
  "symptoms": "string",
  "history": "string"
}

Response:

{
  "answer": "string",
  "clinical_trials": [
    {
      "title": "string",
      "condition": "string", 
      "intervention": "string",
      "eligibility": "string"
    }
  ]
}

2. Index Management

Build Indices

POST /indices/build

Builds the FAISS indices for both medical knowledge and clinical trials.

Response:

{
  "medical_index": "built successfully",
  "clinical_trials_index": "built successfully",
  "clinical_trials_count": 42
}

Check Index Status

GET /indices/status

Returns the current status of all indices.

Response:

{
  "medical_index_built": true,
  "clinical_trials_index_built": true,
  "clinical_trials_count": 42
}

Usage Flow

1. Start the server
2. Call /indices/build to build the knowledge indices
3. Check status with /indices/status if needed
4. Make medical queries using /analyze
5. When asking about clinical trials, include "clinical trials" in your query for automatic suggestions

Important Notes

• Before using the analysis endpoint, make sure to build the indices first
• The system automatically detects "clinical trials" in queries and includes relevant trial suggestions when available