Comprehensive Documentation

Everything you need to understand, deploy, and extend MediChain AI - with complete code examples and implementation details

GitHub Repository

Complete source code, setup instructions, and contribution guidelines

View on GitHub

API Documentation

Agent addresses, message protocols, and integration guides

View Architecture

Quick Start

Test the live demo on Agentverse in under 2 minutes

Try Demo

Code Examples & Implementation

Explore production-ready code snippets from our multi-agent diagnostic system

Message Protocol
python
# Message Protocol Structure (Pydantic Models)
class Symptom(BaseModel):
"""Individual symptom with metadata"""
name: str
raw_text: str
severity: Optional[int] = Field(None, ge=1, le=10)
duration: Optional[str] = None
frequency: Optional[str] = None
class PatientIntakeData(BaseModel):
"""Structured patient symptom data"""
session_id: str
symptoms: List[Symptom]
age: Optional[int] = None
medical_history: Optional[List[str]] = None
allergies: Optional[List[str]] = None
current_medications: Optional[List[str]] = None
2,000+ lines of code
181 tests passing
4 deployed agents

Technical Architecture

Deep dive into system components

Agent Communication Flow

Mailbox Protocol enables async inter-agent messaging through Agentverse infrastructure

uAgents mailbox=True, Pydantic message models, Chat Protocol for ASI:One

MeTTa Knowledge Graph

2,074 medical facts with 34 query methods for transparent diagnostic reasoning

hyperon>=0.1.0, symbolic reasoning, multi-hop queries, evidence tracing

Input Validation System

14 edge case scenarios with safety-first priority (emergency, crisis, boundaries)

Confidence scoring, priority-based validation, flexible NLP detection

Testing Infrastructure

181 comprehensive tests covering all components and medical scenarios

pytest, pytest-asyncio, 84% coverage, zero critical bugs

ASI Alliance Integration

Deep integration with Fetch.ai and SingularityNET

Fetch.aiuAgents Framework

  • •Multi-agent orchestration with coordinator pattern
  • •Mailbox protocol for async communication
  • •Chat Protocol for ASI:One discoverability
  • •Agentverse deployment with 24/7 uptime

SingularityNETMeTTa Knowledge Graph

  • •2,074 medical facts as symbolic knowledge
  • •34 query methods (16 medical-specific)
  • •Transparent reasoning chain generation
  • •Contraindication and drug interaction checks

System Capabilities

Production-ready multi-agent diagnostic system

25
Medical Conditions
2,074
Medical Facts
34
Query Methods
181
Tests Passing
83+
Contraindications
37+
Lab Tests
12
Imaging Types
180
Risk Factors

Resources & Links

Fetch.ai DocumentationMeTTa Language DocsAgentverse PlatformInnovation Lab Examples

Frequently Asked Questions

How accurate is MediChain AI?

Our system achieves 87% diagnostic accuracy on test cases, covering 25 medical conditions with evidence-based reasoning from CDC, WHO, and other authoritative sources.

What medical conditions does it cover?

25 conditions total: 9 critical (meningitis, stroke, MI, PE, appendicitis, anaphylaxis, DKA, sepsis, aortic dissection), 7 urgent (pneumonia, asthma exacerbation, kidney stones, DVT, acute pancreatitis, cholecystitis, diverticulitis), and 9 routine (influenza, UTI, migraine, GERD, allergic rhinitis, tension headache, gastroenteritis, bronchitis, sinusitis).

Is my data stored?

No. All processing happens in real-time via mailbox protocol. No patient data is stored or logged. Complete privacy-first architecture.

How does MeTTa reasoning work?

MeTTa is a symbolic reasoning engine from SingularityNET. It stores 2,074 medical facts as knowledge graphs and performs transparent logical queries to generate diagnosis reasoning chains.

Can I deploy this myself?

Yes! The project is open-source. See our GitHub repository for complete setup instructions, including VPS deployment, Agentverse configuration, and local development.

Need Help?

Check out our GitHub repository or reach out to the team

Open GitHub Issue