How to Configure a Medical Chatbot Using the AI Engine Plugin - part IV

📚 Complete Tutorial: Configuring a Medical Chatbot with AI Engine

*This is the fourth part of the complete configuration guide. If you missed the first three parts, you can find them [here]*, [here]*** and [here]***

In the first three parts we covered:
✅ **Technical Foundation** – Token system, cost optimization, Dashboard modules
✅ **Chatbot Configuration** – AI Model, Embeddings, Thresholds, Knowledge Base
✅ **User Interface** – Appearance, Popup, UI Builder, Advanced
✅ **Advanced Integrations** – Cross-Site, Shortcodes, Actions
✅ **Monitoring** – Discussions, Insights
✅ **Partial Global Settings** – Settings AI

—

## 🎛️ NOW WE WILL FINALIZE THE GLOBAL SETTINGS:

In this final part, we will complete the configuration of all system settings:

🔹 **Settings – Chatbot** – Global chatbot configurations
🔹 **Settings – Knowledge** – Knowledge base settings
🔹 **Settings – Orchestration** – Advanced automations
🔹 **Settings – Files & Media** – File management
🔹 **Settings – Remote Access** – API and MCP
🔹 **Settings – Others** – Security, languages, maintenance

⚙️ “SETTINGS” – GLOBAL CONFIGURATIONS

2. “CHATBOT” TAB – GLOBAL SYSTEM SETTINGS

📍 AI Engine → Settings → “Chatbot” Tab

🔹 “CHATBOT” SECTION – BASIC FUNCTIONALITIES

📝 DISCUSSIONS – ARCHIVING CONVERSATIONS

Status: ❌ Disable (disabled)
Function: Saves and archives all user conversations
Access: "Discussions" Tab for history viewing

🏥 MEDICAL STRATEGY:

✅ DISABLED - Security Rationale:
  - Protection of sensitive medical data
  - Strict GDPR compliance
  - Patient history is not stored
  - Maximum privacy for users

🌐 CROSS-SITE – EXTERNAL EMBED

Status: ✅ Enable (activated)
Function: Allows chatbot embed on external sites
Control: Access based on authorized domains

🚀 IMPLEMENTATION BENEFITS:

Portability: Chatbot can be integrated on multiple platforms
Control: Only authorized domains can use the system
Scalability: Ready for medical partnerships

💻 SYNTAX HIGHLIGHT – CODE DISPLAY

Status: ❌ Disable (disabled)
Function: Syntax highlight for code blocks
Usage: For technical/developer examples

🎯 MEDICAL DECISION:

❌ NOT RELEVANT - Our system:
  - Does not display code or technical examples
  - Purely conversational medical interaction
  - No need for code formatting

🎤 WEB SPEECH API – VOICE RECOGNITION

Speech Recognition: ❌ Disable (disabled)
Function: Speech-to-text conversion
Dependencies: Browser and user OS

🔊 SPEECH SYNTHESIS – TEXT-TO-SPEECH

Status: ❌ Disable (disabled)
Function: Text-to-speech conversion
GDPR: Users must accept the terms

🏥 ACCESSIBILITY ANALYSIS:

✅ DISABLED - Medical Considerations:
  - Critical precision in medical terminology
  - Risks of vocal misunderstanding
  - GDPR complexity for voice data
  - Focus on 100% textual accuracy

🔹 “DISCUSSIONS” SECTION – CONVERSATION MANAGEMENT

📊 SUMMARIZE – AUTOMATIC SUMMARY

Status: ❌ Disable (disabled)
Function: Automatically generates title for discussions
Model: Uses the "fast" model for summaries

🎯 TITLE STRATEGY:

✅ DISABLED - Rationale:
  - Automatic titles can distort medical context
  - We prefer clear titles based on the subject
  - Reduction of unnecessary processing costs

📄 PAGING – RESULT PAGINATION

Value: 10 per Page
Function: Number of discussions displayed per page

⚡ PERFORMANCE OPTIMIZATION:

10 per Page - Perfect Balance:
  - Fast list loading
  - Easy navigation for administrators
  - Optimized for large data volumes

🔄 REFRESH INTERVAL – AUTOMATIC UPDATE

Value: 5 seconds
Function: Automatically updates the list of discussions

📈 REAL-TIME MONITORING:

5 seconds - Optimized for:
  - Live monitoring of conversations
  - Rapid problem detection
  - Prompt intervention if necessary

📋 METADATA BAR – ADDITIONAL INFORMATION

Status: ❌ Disable (disabled)
Available Options:
  - Start Date: discussion creation date
  - Last Update: last modification date
  - Message Count: number of messages

🎯 INTERFACE DECISION:

✅ CLEAN INTERFACE - Priorities:
  - Focus on medical content, not metadata
  - Reduction of visual complexity
  - Simplified user experience

🔹 “USER INTERFACE” SECTION – CHATBOT SELECTION

🎛️ CHATBOT SELECT – SELECTION METHOD

Available Options:
  - [ ] Tabs (default)
  - [✅] Dropdown (selected)
Function: How multiple chatbots are displayed

🏗️ SELECTION ARCHITECTURE:

DROPDOWN SELECTED - Advantages:
  - Scalable for many chatbots
  - Compact and organized interface
  - Easy search and filtering
  - Ideal for future expansions

🏥 GLOBAL CHATBOT SETTINGS STRATEGY

🔒 MEDICAL SECURITY AND COMPLIANCE

📜 DATA POLICY:

Discussions Disabled:
  - No medical conversations stored
  - Maximum patient privacy protection
  - Zero risk of sensitive data breach

Speech APIs Disabled:
  - Eliminates risks of vocal interpretation
  - Strict GDPR compliance
  - Focus on textual accuracy

🌐 SCALABILITY AND PORTABILITY

🚀 MULTI-SITE ARCHITECTURE:

Cross-Site Enabled:
  - Ready for integration on multiple platforms
  - Granular control over authorized domains
  - Expansion without reconfiguration

Dropdown Interface:
  - Support for future specialized chatbots
  - Clear organization of multiple medical services

⚡ PERFORMANCE AND EXPERIENCE

🎨 OPTIMIZED INTERFACE:

Clean Interface:
  - No unnecessary elements for the medical use-case
  - Focus on essential functionalities
  - Fast and responsive loading

Rapid Refresh:
  - Real-time activity monitoring
  - Immediate intervention if necessary

✅ FINAL RECOMMENDED CONFIGURATION

🎛️ OPTIMAL SETTINGS FOR MEDICAL SYSTEM:

# CHATBOT SECTION
Discussions: ❌ Disable (data security)
Cross-Site: ✅ Enable (scalability)
Syntax Highlight: ❌ Disable (not relevant)
Speech Recognition: ❌ Disable (GDPR)
Speech Synthesis: ❌ Disable (precision)

# DISCUSSIONS SECTION
Summarize: ❌ Disable (accuracy)
Paging: 10 per Page (performance)
Refresh: 5 seconds (monitoring)
Metadata Bar: ❌ Disable (simplicity)

# USER INTERFACE SECTION
Chatbot Select: ✅ Dropdown (scalability)

3. “KNOWLEDGE” TAB – KNOWLEDGE BASE CONFIGURATION

📍 AI Engine → Settings → “KNOWLEDGE” Tab

🔹 PANEL 1: “ENVIRONMENT FOR EMBEDDINGS” – CREATING CONNECTIONS

This panel allows creating new connections to vector databases.

📝 NAME – CUSTOM NAME:

Value: "Quadrant" (custom example)
Function: Descriptive name for your connection
Usage: Easy identification in the list
Examples: "MedicalBase", "PrimaryEmbeddings", etc.

🔧 TYPE – VECTOR DATABASE PROVIDER TYPE:

Available Options:
├─ Qdrant (✅ Selected in example)
├─ Pinecone
├─ OpenAI Vector Store
└─ Chroma

📊 PROVIDER COMPARISON:

| Provider | Cost | Performance | Ease of Use | Use-case |
|------------|------|---------------|------------|---------------|
| Qdrant | 🟢 Free 1GB | 🟢 Excellent | 🟢 Easy | Small-medium projects |
| Pinecone | 🔵 Pay-per-use | 🟢 Very Good | 🟢 Easy | Enterprise projects |
| OpenAI | 🔵 Integrated | 🟡 Average | 🟢 Simple | OpenAI only |
| Chroma | 🟢 Open-source | 🟡 Good | 🟡 Medium | Self-hosted |

🌍 SERVER – CONNECTION URL:

Example: "https://a4c16481-67f9-4932-9166-ef4055ed00c2.europe-west3-0.gcp.cloud.qdrant.io"
Source: Obtained from the Qdrant Cloud platform
Structure: [unique-ID].[region].[provider].cloud.qdrant.io

🗂️ COLLECTION – DATA COLLECTION:

Value: "1536" (default)
Function: Separates data into logical "compartments"
Analogous: Like folders on a hard drive
Usage: You can have multiple collections for different projects

📐 DIMENSIONS – VECTOR DIMENSIONS:

Value: 1536 (default for Ada-002)
Significance: The number of characteristics of each vector
Compatibility: Must match the embeddings model

🆔 ENV ID – AUTOMATIC IDENTIFIER:

Example: "unynjwo1" (automatically generated)
Function: Unique ID for the connection
Usage: Technical references in the backend

🎯 SCORE – ADVANCED SEARCH SETTINGS:

Min Score: 70 (default)
Significance: The minimum similarity threshold for results
Example: Only vectors with >70% similarity are returned

Max Results: 5 (default)
Significance: The maximum number of results returned
Interval: 1-100 (10 is general default)

⚙️ AI ENVIRONMENT – MODEL CONFIGURATION:

Override Defaults: ❌ Disable (unchecked)
Function: Allows custom settings different from global ones
Usage: Only if you need a different model for this environment

🔧 ACTIONS – PRACTICAL ACTIONS:

Quick Test: ✅ Connection test button
Function: Checks if the connection is working
Delete: 🗑️ Environment deletion button
Caution: Permanently deletes the connection and all data

🏗️ QDRANT CONNECTION ARCHITECTURE

🔗 COMPLETE INTEGRATION FLOW:

AI Engine → Qdrant Cloud → Collection 1536 → Medical Vectors
    ↓           ↓               ↓              ↓
Connection   GCP Cluster   Data Separation    Medicine Embeddings
API Key   Server URL    1536 Dimensions    Cosine Similarity

💊 BENEFITS FOR THE MEDICAL SYSTEM:

✅ Performance: Ultra-fast vector search in the medical database
✅ Scalability: Qdrant Cloud supports large volumes of medicines
✅ Security: Secure connection via HTTPS with API Key authentication
✅ Cost: Optimized with free plan (1GB) and 5% discount with #AIENGINES

🔍 COMPATIBILITY WITH EMBEDDING ADA-002:

Dimensions: 1536 ←→ 1536 (perfect match)
Similarity: Cosine ←→ Cosine (optimal for medical text)
Collection: Auto-creation ←→ Native dimensions

🔹 PANEL 2: “DEFAULT EMBEDDINGS ENVIRONMENT” – SELECTING ACTIVE CONNECTION

Here you choose which environment should be used by default.

Options: List of all created environments
Selection: "Quadrant" (our created environment)
Impact: All searches use this connection by default

🔹 PANEL 3: “DEFAULTS” – GLOBAL CONFIGURATION

Global settings for all environments (if not overridden).

AI Environment: OpenAI (the main provider)
Embeddings Model: text-embedding-ada-002 (the optimal model)
Dimensions: 1536 (Ada-002 native dimensions)

Note: "If a particular embeddings environment needs different settings,
use the 'Override Defaults' option in the AI Environment section
of each embeddings environment."

🔹 PANEL 4: “EMBEDDINGS SEARCH” – SEARCH STRATEGY

Configuring how the system searches the database.

🔍 METHOD – SEARCH METHOD:

Available Options:
├─ Simple (✅ Selected)
├─ Context-Aware
└─ Smart Search

📊 METHOD ANALYSIS:

Method          Speed          Accuracy        Cost          Usage
───────         ───────        ─────────        ───────        ───────────────
Simple          🟢 Fast        🟡 Average      🟢 Free        Simple questions

Context-Aware   🟡 Average      🟢 Good         🟢 Free        Medical conversations

Smart Search    🔵 Slow        🟢 Excellent     🔵 Added Cost  Complex searches

💬 MESSAGES – NUMBER OF CONTEXT MESSAGES:

Value: 10 (default)
Function: How many messages from history are used for context
Impact: The more messages, the richer the context
Recommendation: 5-10 for a balance between context and performance

🔹 PANEL 5: “INFORMATION” – THEORETICAL EXPLANATIONS

Informational panel explaining the technical concepts.

🎓 WHAT EMBEDDINGS ARE:

Definition: "Textual data converted into vectors that enable similarity search"
Function: "Allow AI to find relevant context from your knowledge base"
Storage: "Synchronized with vector databases like Pinecone or Qdrant"

🛠️ WORKING WITH EMBEDDINGS – PRACTICAL FLOW:

Knowledge Tab: "Manage your embeddings"
Available Actions:
├─ EDIT: "Create, edit, and search embeddings"
├─ AI SEARCH: "Query your knowledge base directly"
└─ SYNC: "Process posts and create/update embeddings"

🚀 MEDICAL CONFIGURATION RECOMMENDATIONS:

✅ Environment Type: Qdrant (zero cost, good performance)
✅ Search Method: Context-Aware ( ideal for medical conversations)
✅ Min Score: 70 (balance between precision and coverage)
✅ Max Results: 5-10 (sufficient for medical recommendations)
✅ Messages: 10 (adequate context for medical history)

4. “ORCHESTRATION” TAB – MODEL CONTEXT PROTOCOL (MCP)

📍 AI Engine → Settings → Orchestration

🔹 WHAT IS MCP (MODEL CONTEXT PROTOCOL)?

🎯 TECHNICAL DEFINITION:

MCP: "Model Context Protocol" - open-source standard created by Anthropic
Purpose: Allows AI models to access external tools and data in real time
Analogous: "Plugin system" for AI - like browser extensions

**MCP ARCHITECTURE IN AI ENGINE:**

AI Engine       →       MCP Client       →       External MCP Server       →       Specialized Tools
     ↓                       ↓                           ↓                             ↓
Medical Chatbot      Standard Protocol        External Services              Functions Available for AI
                     (communication)            (APIs, databases)              (search, calculations, etc.)

🔹 WHAT DO MCP SERVERS DO?

🛠️ PRACTICAL FUNCTIONALITIES:

MCP Servers offer:
- Web Search - real-time information retrieval
- Calculator - complex mathematical calculations
- Weather - current weather data
- Calendar - calendar interaction
- Database Query - database querying
- File System - file reading/writing
- Custom Tools - personalized business functions

💊 CONCRETE MEDICAL EXAMPLE:

User              →       Chatbot          →       MCP Server          →       Medical Database
     ↓                       ↓                           ↓                             ↓
"Drug Interaction    Detects the need        Calls the DrugInteractions    Queries the medical
 Warfarin +            for external medical              API                         database
 Amoxicillin?"          information

Response: "Warfarin + Amoxicillin can increase the risk of bleeding"

🔹 CONFIGURING THE MCP SERVER

**👤 USER NAME – UNIQUE NAME:**

Example: "DrugInteractionsAPI"
Requirement: "The name must be unique across all MCP servers"
Usage: Visual identification in the list

**🌐 URL – MCP SERVER ENDPOINT:**

Structure: "https://example.com/mcp-server"
Protocol: "Must start with https://" (mandatory security)
Requirements: Must implement the standard MCP protocol

**🔐 TOKEN – AUTHENTICATION:**

Type: "Bearer token for authentication"
Function: Access token for the MCP server
Security: Ensures only authorized users

**🆔 SERVER ID – IDENTIFIER:**

Example: "7x86fkqk" (automatically generated)
Purpose: Unique identification in the system

**🔧 ACTIONS – AVAILABLE ACTIONS:**

Delete: Deletes the MCP server
Caution: Deletion is permanent and irreversible

🔹 WHY “ORCHESTRATION”?

**THE ORCHESTRA ANALOGY:**

Orchestrator        →       Instruments       →       Score       →       Result
     ↓                       ↓                       ↓                       ↓
AI Engine           MCP Servers              User Conversation           Coordinated Harmonious
(the conductor)           (the musical              (the music)                 Response (the symphony)
                    instruments)

**COMPLETE ORCHESTRATION FLOW:**

1. User           →       2. AI Engine       →       3. Orchestration       →       4. MCP Server
     ↓                       ↓                           ↓                             ↓
"What's the weather      Detects the need        Calls the "Weather"          Queries the weather
 like in Bucharest?"      for weather data             MCP Server                   API

5. AI Engine        →       6. User
     ↓                       ↓
Integrates the response     Receives "In Bucharest
into the conversation       it's 22°C, sunny"

🏥 SPECIFIC MEDICAL APPLICATIONS

**MCP SERVERS FOR MEDICINE:**

- Drug Databases: Updated drug databases
- Medical Calculators: Dosage calculators, BMI, etc.
- Clinical Guidelines: Current medical guidelines
- Lab Reference: Normal analysis values
- Symptom Checkers: Symptom triage systems

**LIMITATIONS AND MEDICAL CONSIDERATIONS:**

- Does not replace medical consultation
- No access to patients' personal data
- Only general medical information
- Mandatory verification of credible sources
- Mention "General information, consult a doctor"

🔒 SECURITY AND COMPLIANCE

**PROTECTION MEASURES:**

- Mandatory HTTPS: All connections encrypted
- Token Authentication: Controlled access
- Server Validation: Verification of secure sources
- Data Minimization: Only necessary data

**✅ CONFIGURATION RECOMMENDATIONS:**

Current Status: Empty panel (no MCP servers configured)
Recommendation: Kept this way for maximum security
Reason: The medical system functions perfectly with its own database

5. “FILES & MEDIA” TAB – FILE AND MEDIA MANAGEMENT

📍 AI Engine → Settings → Files & Media

📤 UPLOADED BY USERS – FILES UPLOADED BY USERS

**🔹 LOCAL UPLOAD – Configuration:**

Storage: Filesystem (✅ Selected)
Options:
  ├─ Filesystem
  └─ Media Library

**🔹 REMOTE UPLOAD – Configuration:**

Storage: Filesystem (✅ Selected)
Options:
  ├─ Filesystem
  └─ Media Library
Note: "Files can be stored either in the filesystem or the Media Library"

**🔹 EXPIRATION – Expiration Policy:**

Setting: 1 hour (✅ Selected)
Options available:
  ├─ 5 minutes
  ├─ 1 hour
  ├─ 6 hours
  ├─ 1 day
  ├─ 1 week
  ├─ 1 month
  └─ Never

**📝 EXPLANATORY NOTE:**

"Uploaded files will be deleted after a certain amount of time.
This also affects files uploaded to OpenAI via the Assistants."

🤖 GENERATED BY AI – FILES GENERATED BY AI

**🔹 LOCAL DOWNLOAD – Configuration:**

Storage: None (✅ Selected)
Options:
  ├─ None
  ├─ Filesystem
  └─ Media Library

**🔹 EXPIRATION – Expiration Policy:**

Setting: Never (✅ Selected)
Note: "Downloaded files will be deleted after a certain amount of time"

🏗️ TECHNICAL CHOICE ARCHITECTURE

**🎯 SELECTED STORAGE STRATEGY:**

**✅ UPLOADED FILES → Filesystem:**

Motivation:
  - Clear separation from the WordPress Media Library
  - Granular control over temporary files
  - Ease of automatic cleanup

**✅ GENERATED FILES → None:**

Motivation:
  - The medical chatbot does not generate files
  - Eliminates unnecessary storage overhead
  - Focus on pure text conversation

**✅ EXPIRATION → 1 hour / Never:**

Uploaded: 1 hour →
  - User files are temporary
  - Automatic cleanup for space saving
  - Alignment with the episodic nature of medical consultations

Generated: Never →
  - Not applicable (storage set to None)
  - Consistency with the decision not to store generated files

🔒 IMPLICATIONS FOR MEDICAL SECURITY

**🛡️ PATIENT DATA PROTECTION:**

Rapid expiration (1 hour):
  - Minimizes the risk of sensitive data exposure
  - Files are not stored long-term
  - Compliance with data protection regulations

Separate Storage:
  - Isolation from the public WordPress Media Library
  - Better control over file access

⚠️ IMPORTANT! CRITICAL CONNECTION WITH CHATBOT CONFIGURATION

MANDATORY CONDITION:
├─ For "Files & Media" settings to have any effect
├─ "File Upload" must be enabled in the chatbot configuration
├─ Tab: AI Engine → Chatbots → [Chatbot Name] → "AI Model" Tab
└─ Setting: Uploads → File Upload ✅ ENABLED

⚠️ CONFIGURATION ALERT:
"Without enabling File Upload in the AI Model tab,
all settings in Files & Media are INACTIVE!"

💊 IMPACT ON THE MEDICAL SYSTEM

**✅ BENEFITS OF OUR CONFIGURATION:**

Security:
  - User files expire quickly (1 hour)
  - Zero generated files permanently stored
  - Separation from public WordPress content

Performance:
  - Optimized storage space
  - Automatic cleanup of temporary resources
  - No overhead for generated files

Compliance:
  - Minimization of sensitive data stored
  - Alignment with data minimization principles
  - Protection against personal data accumulation

**🎯 CONCLUSION OF FILES & MEDIA CONFIGURATION:**

The chosen settings create a perfect balance between:

**Security** (rapid expiration + separate storage)
**Performance** (no unnecessary storage)
**Compliance** (data minimization)

But everything depends on enabling **File Upload** in the chatbot configuration!

6. “REMOTE ACCESS” TAB – REMOTE ACCESS AND INTEGRATIONS

📍 AI Engine → Settings → Remote Access

🔹 “MODEL CONTEXT PROTOCOL (MCP)” PANEL – WORDPRESS AS MCP SERVER

**🎯 TRANSFORMING WORDPRESS INTO AN MCP SERVER:**

Function: "Create a Model Context Protocol (MCP) server that provides various tools"
Purpose: Makes the WordPress site accessible as a service for external AI agents
Direction: ← AS (WordPress becomes the server for others)

**⚙️ SSE ENDPOINT CONFIGURATION:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Enable the App.json/msg/v1/spe endpoint"
Note: "Claude works well with MCP servers but cannot handle SSE directly, so a relay is needed"

**🔐 BEARER TOKEN – SECURITY:**

Setting: [Empty]
Function: "MCP will be usable by using this Bearer Token"
Security: If not set, it requires personalized authentication filters

**🌐 NO-AUTH URL – SIMPLIFIED ACCESS:**

Setting: [ ] Enable (❌ Disabled)
Condition: "Only works when SSE Endpoint is enabled and Bearer Token is set"
Function: Creates a special URL with authentication in the path

🔹 “REST API” PANEL – AUTOMATIONS AND INTEGRATIONS

**🎯 API FOR EXTERNAL AUTOMATIONS:**

Definition: "The REST API enables integration with automation platforms and external applications"
Usage: "Use it with Make.com, n8n, Zapier, or any platform that supports REST APIs"

**⚙️ PUBLIC API – ACTIVATION:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Enable the Public REST API"

**🔐 BEARER TOKEN – API SECURITY:**

Setting: [Empty]
Function: "The Public API will be usable by using this Bearer Token"
Note: "If not set, you will need to build your own authentication"

🔹 “INFORMATION” PANEL – TECHNICAL EXPLANATIONS

**🔄 INTEGRATION COMPARISON:**

REST API:
├─ Purpose: "Ideal for automation platforms like Make.com, Zapier, or n8n"
├─ Usage: "Trigger AI operations from external services"
└─ Cases: Integration into existing workflows

MCP (Model Context Protocol):
├─ Purpose: Makes WordPress accessible for AI agents
├─ Compatibility:
│  ├─ ✅ Claude Desktop: "Use the mcp.js relay"
│  ├─ ⚠️ OpenAI/ChatGPT: "Limited to Deep Research mode"
│  └─ ❌ Claude.ai: "Currently not supported"
└─ Available Tools:
   ├─ Full access: WordPress management, SEO Engine, Code Engine
   └─ Limited access: Only search and fetch WordPress posts

**⚡ CRITICAL DIRECTION NOTE:**

"Note: This makes WordPress an MCP server.
For connecting TO external MCP servers, use the Orchestration module."

📋 FINAL COMPARISON:

| Direction | Purpose | Module | Location |
|------------|------------------------|---------|------------------------|
| → TO | Chatbot calls external services | MCP Client | Settings → Orchestration |
| ← AS | WordPress becomes a server for AI agents | MCP Server | Settings → Remote Access |

🏗️ TECHNICAL DECISION ARCHITECTURE

**🎯 SELECTED SECURITY STRATEGY:**

**✅ SSE ENDPOINT → DISABLED:**

Medical Motivation:
  - The medical system does not require external access to WordPress
  - Reduces the attack surface and vulnerabilities
  - Keeps medical data isolated and secured

**✅ REST API → DISABLED:**

Medical Motivation:
  - The medical chatbot operates autonomously, without external integration
  - Eliminates the risk of data exposure via API
  - Simplifies the architecture and reduces points of failure

**✅ BEARER TOKENS → UNSET:**

Medical Motivation:
  - No API keys = no external access
  - Maximum protection against unauthorized access
  - Alignment with medical security principles

🔒 IMPLICATIONS FOR MEDICAL SECURITY

**🛡️ BENEFITS OF COMPLETE ISOLATION:**

Data Protection:
  - Zero external access to the medical system
  - Patient data remains completely internal
  - No exposed public endpoints

Compliance:
  - Maximized GDPR/HIPAA compliance
  - Total control over data flow
  - Simplified and secured audit trail

Stability:
  - System independent of external services
  - No dependencies on external connections
  - Consistent and predictable performance

⚠️ IMPORTANT! RELATIONSHIP WITH ORCHESTRATION

CRITICAL DISTINCTION:

ORCHESTRATION (Tab 4):                    REMOTE ACCESS (Tab 6):
├─ Direction: → TO                        ├─ Direction: ← AS
├─ WordPress = Client                    ├─ WordPress = Server
├─ Calls external services              ├─ Is called by external agents
├─ Extends capabilities                    ├─ Exposes capabilities
└─ Used for MCP Servers            └─ Used to become an MCP Server

🚨 CONSCIOUS DECISION:
"Both modules are DISABLED to keep
the medical system completely isolated and secured!"

**🎯 CONCLUSION OF REMOTE ACCESS CONFIGURATION:**

The chosen configuration ensures **complete isolation** of the medical system:

**No inbound access** (Remote Access disabled)
**No outbound access** (Orchestration empty)
**Closed and secured ecosystem** for sensitive medical data

The system functions perfectly with its internal database, without external dependencies! 🔒

7. “OTHERS” TAB – ADVANCED AND MISCELLANEOUS SETTINGS

📍 AI Engine → Settings → Others

🔹 PANEL 1 – “USER INTERFACE” – USER INTERFACE

**🎯 USER EXPERIENCE CONFIGURATION:**

**📝 INTRO MESSAGE:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Enable introduction messages on some modules"
Medical Motivation:
  - The medical chatbot is direct and professional
  - Eliminates unnecessary introductory messages for consultations
  - Fast focus on the user's medical issue

**📋 FORMS EDITOR:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Enable the Forms module to use the Forms Editor"
Medical Motivation:
  - The medical system does not require complex forms
  - Pure conversational interaction is sufficient
  Simplifies: the interface and reduces overhead

🔹 PANEL 2 – “INSIGHTS” – ANALYTICS AND MONITORING

**🎯 ACTIVITY MONITORING POLICY:**

**📊 QUERIES DATA:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Record the queries and their replies"
Note: "Contextual data is excluded unless you activate the options below"
Medical Motivation:
  - Maximum protection of patient confidentiality
  - Compliance with GDPR - minimization of stored data
  - Eliminates the risk of exposure of medical consultations

**📝 QUERIES FORM DATA:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Capture form data neatly for simple reuse"
Medical Motivation:
  - Medical data should not be captured for reuse
  - Each consultation is unique and personal
  - Respects the principle of medical confidentiality

**⚠️ AWARENESS NOTE:**

"Make sure your users are aware of this"
Decision: Users are automatically protected by disabling

🔹 PANEL 3 – “ADMIN BAR” – ADMINISTRATION BAR

**🎯 QUICK ACCESS FOR ADMINS:**

**🤖 AI ENGINE:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Medical Motivation:
  - Reduces clutter in the administration interface
  - Access to settings is done directly from the plugin
  - Keeps the interface clean and professional

**🎮 PLAYGROUND:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Medical Motivation:
  - The medical system is not a "playground"
  - Eliminates the temptation to experiment in production
  - Maintains focus on serious medical functionalities

**📝 GENERATE CONTENT:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Medical Motivation:
  - The medical chatbot does not generate generic content
  - Exclusive specialization in medical consultations
  - Prevents improper use of resources

**🖼️ GENERATE IMAGES:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Medical Motivation:
  - The medical system does not require image generation
  - Exclusive focus on text-based interaction
  - Saves processing resources

🔹 PANEL 4 – “ADVANCED” – ADVANCED SETTINGS

**🎯 ADVANCED TECHNICAL CONFIGURATIONS:**

**🔣 SHORTCODES:**

Setting: [ ] Resolve (❌ Disabled - RECOMMENDED)
Function: "Will resolve the shortcode in your content when needed"
Medical Motivation:
  - The medical chatbot does not process shortcodes
  - Direct interaction with the drug database
  - Eliminates unnecessary complexity

**📏 CONTEXT MAX LENGTH:**

Setting: 16384 (✅ Kept default)
Medical Motivation:
  - Perfect balance between context and performance
  - Sufficient for detailed medical conversations
  - Optimized for cost and response speed

**🔧 DEV TOOLS:**

Setting: [ ] Enable (❌ Disabled - RECOMMENDED)
Function: "Enable a new tab with debugging tools. For developers only"
Medical Motivation:
  - The medical system in production does not require debugging
  - Increased security by disabling the tools
  - Reduces the risk of unauthorized interventions

**🗑️ PLUGIN DATA:**

Setting: [ ] Delete All (❌ Disabled - NORMAL)
Function: "The database and all the options will be removed on uninstall"
Medical Motivation:
  - Safety setting - not activated preemptively
  - Retains important configuration data
  - Conscious decision at the time of uninstallation

🔹 PANEL 5 – “LANGUAGES” – FOREIGN LANGUAGE SUPPORT

**🎯 MULTILINGUAL CONFIGURATION:**

**🌐 AVAILABLE LANGUAGES:**

Setting: English (en) (✅ Kept default)
List: English (en), German (de), French (fr), Spanish (es), Italian (it), Chinese (zh), Japanese (ja), Po
Medical Motivation:
  - The Romanian medical chatbot works in English
  - The drug database is in Romanian/English
  - Eliminates the complexity of multi-language support

🔹 PANEL 6 – “SECURITY” – ADVANCED SECURITY

**🎯 ADVANCED PROTECTION MEASURES:**

**🔒 PRIVACY FIRST:**

Setting: [✓] Enable (✅ ENABLED - CRITICAL)
Function: "Only essential data is stored. IPs, conversations, and other private details are minimized"
Medical Motivation:
  - Maximum protection of patient confidentiality
  - Compliance with medical regulations (GDPR/HIPAA)
  - Minimization of sensitive data stored

**🚫 BANNED WORDS:**

Setting: [Empty Field] (✅ Kept empty - RECOMMENDED)
Function: "Reject queries containing these words; separate with commas"
Medical Motivation:
  - The medical chatbot must respond to any medical question
  - Overly aggressive filters can block legitimate medical questions
  - The system relies on strict prompts, not word filters

**🛡️ BANNED IPS:**

Setting: [Empty Field] (✅ Kept empty - NORMAL)
Function: "Block access from specified IP addresses or ranges using CIDR notation"
Medical Motivation:
  - The medical system is public and accessible to everyone
  - IP restrictions are not necessary at this time
  - Can be configured later if the need arises

🔹 PANEL 7 – “MAINTENANCE” – MAINTENANCE AND BACKUP

**🎯 CONFIGURATION MANAGEMENT:**

**💾 SETTINGS BACKUP:**

Export Settings: (✅ Available - IMPORTANT)
Import Settings: (✅ Available - IMPORTANT)
Reset Settings: (✅ Available - CAUTIOUS)
Reset Usage: (✅ Available - USEFUL)

Medical Strategy:
  - Periodic export of configuration for backup
  - Rapid import in case of need
  - Reset usage for current monitoring

**📝 MAINTENANCE NOTE:**

"Important to keep regular backups of your settings"
Recommended Practice: Weekly export of the complete configuration

🏗️ GENERAL STRATEGY FOR “OTHERS”

**🎯 CONFIGURATION PHILOSOPHY:**

Functional Minimalism:
  - Only strictly necessary settings are enabled
  - Maximized security and privacy
  - Complete elimination of unnecessary functionalities

Medical Focus:
  - All decisions justified by medical necessity
  - Patient data protection is prioritized
  Simplification: for maximum reliability

**🔒 GLOBAL SECURITY ENSURED:**

Privacy First: ✅ ENABLED - Maximum data protection
Insights: ❌ DISABLED - Zero conversation tracking
Admin Features: ❌ DISABLED - Minimal interface
Banned Words: ❌ DISABLED - Unlimited medical access

**✅ FINAL CONCLUSION OF “OTHERS” TAB:**

All 7 panels are configured for:

**Maximum security** (Privacy First enabled, Insights disabled)
**Operational simplicity** (unnecessary features disabled)
**Exclusive medical focus** (only relevant functionalities)
**Complete compliance** with medical standards

The system is now optimized for its specific medical mission! 🏥

**📚 CONTINUING IN PART 5 (FINAL)**

We have finalized the configuration of all global system settings.
In the last article of the series, we will explore the crucial **KNOWLEDGE** tab:

🔹 **Knowledge** – Centralized management of medical databases

Configuration and synchronization of knowledge bases
Optimizations for efficient semantic search
Maintenance and update strategies

[👉 Read Part 5: Knowledge Base Management here]

Recent Posts

Recent Comments

How to Configure a Medical Chatbot Using the AI Engine Plugin – part IV

⚙️ “SETTINGS” – GLOBAL CONFIGURATIONS

2. “CHATBOT” TAB – GLOBAL SYSTEM SETTINGS

🔹 “CHATBOT” SECTION – BASIC FUNCTIONALITIES

🔹 “DISCUSSIONS” SECTION – CONVERSATION MANAGEMENT

🔹 “USER INTERFACE” SECTION – CHATBOT SELECTION

🏥 GLOBAL CHATBOT SETTINGS STRATEGY

🔒 MEDICAL SECURITY AND COMPLIANCE

🌐 SCALABILITY AND PORTABILITY

⚡ PERFORMANCE AND EXPERIENCE

✅ FINAL RECOMMENDED CONFIGURATION

3. “KNOWLEDGE” TAB – KNOWLEDGE BASE CONFIGURATION

🔹 PANEL 1: “ENVIRONMENT FOR EMBEDDINGS” – CREATING CONNECTIONS

🏗️ QDRANT CONNECTION ARCHITECTURE

🔹 PANEL 2: “DEFAULT EMBEDDINGS ENVIRONMENT” – SELECTING ACTIVE CONNECTION

🔹 PANEL 3: “DEFAULTS” – GLOBAL CONFIGURATION

🔹 PANEL 4: “EMBEDDINGS SEARCH” – SEARCH STRATEGY

🔹 PANEL 5: “INFORMATION” – THEORETICAL EXPLANATIONS

4. “ORCHESTRATION” TAB – MODEL CONTEXT PROTOCOL (MCP)

🔹 WHAT IS MCP (MODEL CONTEXT PROTOCOL)?

🔹 WHAT DO MCP SERVERS DO?

🔹 CONFIGURING THE MCP SERVER

🔹 WHY “ORCHESTRATION”?

🏥 SPECIFIC MEDICAL APPLICATIONS

🔒 SECURITY AND COMPLIANCE

5. “FILES & MEDIA” TAB – FILE AND MEDIA MANAGEMENT

📤 UPLOADED BY USERS – FILES UPLOADED BY USERS

🤖 GENERATED BY AI – FILES GENERATED BY AI

🏗️ TECHNICAL CHOICE ARCHITECTURE

🔒 IMPLICATIONS FOR MEDICAL SECURITY

⚠️ IMPORTANT! CRITICAL CONNECTION WITH CHATBOT CONFIGURATION

💊 IMPACT ON THE MEDICAL SYSTEM

6. “REMOTE ACCESS” TAB – REMOTE ACCESS AND INTEGRATIONS

🔹 “MODEL CONTEXT PROTOCOL (MCP)” PANEL – WORDPRESS AS MCP SERVER

🔹 “REST API” PANEL – AUTOMATIONS AND INTEGRATIONS

🔹 “INFORMATION” PANEL – TECHNICAL EXPLANATIONS

🏗️ TECHNICAL DECISION ARCHITECTURE

🔒 IMPLICATIONS FOR MEDICAL SECURITY

⚠️ **IMPORTANT! RELATIONSHIP WITH ORCHESTRATION**

7. “OTHERS” TAB – ADVANCED AND MISCELLANEOUS SETTINGS

🔹 PANEL 1 – “USER INTERFACE” – USER INTERFACE

🔹 PANEL 2 – “INSIGHTS” – ANALYTICS AND MONITORING

🔹 PANEL 3 – “ADMIN BAR” – ADMINISTRATION BAR

🔹 PANEL 4 – “ADVANCED” – ADVANCED SETTINGS

🔹 PANEL 5 – “LANGUAGES” – FOREIGN LANGUAGE SUPPORT

🔹 PANEL 6 – “SECURITY” – ADVANCED SECURITY

🔹 PANEL 7 – “MAINTENANCE” – MAINTENANCE AND BACKUP

🏗️ GENERAL STRATEGY FOR “OTHERS”

Leave a comment Cancel reply

Policies

⚠️ IMPORTANT! RELATIONSHIP WITH ORCHESTRATION