Skip to content

ASHAI Documentation

Welcome to ASHAI - a comprehensive AI Public Health Agent that provides evidence-based medical information through multiple AI-powered endpoints.

What is ASHAI?

ASHAI is a medical AI platform built on PubMed research and curated FAQ sources that offers:

  • Evidence-Based Responses - Strictly based on medical literature and curated sources
  • Intelligent Source Routing - Automatically selects the best knowledge sources for each query
  • Cultural Sensitivity - Provides appropriate guidance for diverse populations
  • Quality Assurance - Built-in evaluation and quality monitoring

Key Components

🤖 AI Agents

Conversational agents that interact with patients and provide medical guidance

Learn about Agents → | Try the Agents →

🎙️ User Interfaces

Voice chat and web interfaces for interacting with ASHAI

Explore Voice Interfaces →

🔍 Search Systems

Multiple specialized medical knowledge sources and databases

Try the Search →

📈 Evaluation Framework

Quality assessment and performance monitoring tools

View Evaluation Documentation → | See Live Leaderboard →

⚙️ API Reference

Interactive FastAPI documentation with full request/response schemas

Open Live API Docs (Swagger) →

Quick Start

Get ASHAI running in 3 simple steps:

  1. Clone and Setup 

    git clone https://github.com/your-org/ashai
cd ashai
source venv/bin/activate
pip install -r requirements.txt

  2. Configure Environment 

    export OPENAI_API_KEY="your_openai_api_key"
# Optional: Google Cloud Healthcare NLP
export GOOGLE_CLOUD_PROJECT_ID="your-project-id"

  3. Start the Server 

    ./run.sh

Visit http://localhost:8000 to access the web interface.

🚀 ← Back to ASHAI Application

Once you've read the documentation, try the live application with all the agents and features described here!

First Time Setup

Check the Quick Start Guide for detailed setup instructions including Google Cloud Healthcare NLP configuration.

Architecture Overview

ASHAI uses a modular architecture with intelligent routing:

graph TB
    User[User Query] --> Agent[ASHAI Agent]
    Agent --> Router[Intelligent Router]
    Router --> |Pregnancy| Googlesheet[Googlesheet FAQ]
    Router --> |General Medical| PubMed[PubMed Search]
    Router --> |Structured Content| September[September Health Library]

    Googlesheet --> Evaluator[Quality Evaluator]
    PubMed --> Evaluator
    September --> Evaluator

    Evaluator --> Response[Evidence-Based Response]

Core Features

🧠 Intelligent FAQ Routing

  • Pregnancy Queries → Googlesheet FAQ + PubMed
  • General Medical → September Health Library + PubMed
  • Regional Health → Bangladesh FAQ + PubMed

Using multiple specialized databases: - Query structured medical content from September Health Library - Search peer-reviewed research via PubMed - Access pregnancy-specific knowledge from Googlesheet FAQ

🛡️ Quality Assurance

  • Evidence-Only Responses - No AI hallucination, only sourced information
  • Automatic Evaluation - Multi-dimensional quality scoring
  • Source Transparency - All sources cited and linked
  • Performance Monitoring - Comprehensive logging and analytics

Example Use Cases

Pregnancy Nutrition

Query: "Can I eat fish during pregnancy?"

Routing: Googlesheet FAQ + PubMed
Response: Evidence-based guidance with specific recommendations and sources

Medication Interactions

Query: "Interactions between diabetes medications and blood pressure drugs"

Routing: PubMed + September Health Library Response: Research evidence with structured medical content

General Health

Query: "What are symptoms of high blood pressure?"

Routing: September Health Library + PubMed
Response: Comprehensive symptom information from authoritative sources

API Endpoints

API Documentation

Getting Help

ASHAI is designed for healthcare professionals and researchers. Always consult qualified medical professionals for medical decisions.