NetScore AI Hub — Medha Admin

Active Tenants

Products

Active Users

Active Documents

Tokens Today

Queries Today

Avg Response

Open Alerts

🧠 LLM Infrastructure Health

🖥️ Ollama Active

Models:

Click Refresh to check

⚡ vLLM Active

Models:

Click Refresh to check

🌐 OpenAI Active

Model:

No API key configured — failover disabled

Click Refresh to check

Current model: Temperature: Max tokens:

Name	Slug	Products	Users	Tokens/Month	Quota	Status	Actions

No tenants found

⏳ user(s) awaiting approval

⏳ Pending Approval Requests

Name	Email	Registered	Email Verified	Note	Actions

Name	Email	Role	Tenant	Status	Last Login	Actions
		📤 Upload 🔄 Reindex
No users found

Name	Code	Collection	Active Docs	Tenants	Status	Actions

Role Name	Scope	Type	Users	Actions

🔐 Role Access Matrix

✅ = Full Access 👁 = View Only ❌ = No Access

Feature / Section	Super Admin Full platform	Core Admin Platform ops	Product Admin Assigned products	Tenant Admin Own tenant	User Chat only
Admin Dashboard
Overview & Stats	✅	✅	👁 assigned only	👁 own tenant	❌
LLM Health Monitor	✅	✅	👁	❌	❌
Kill Switch (AI on/off)	✅	❌	❌	❌	❌
User Management
View all users	✅	✅	❌	👁	❌
Create / Edit users	✅	✅	❌	❌	❌
Reset user passwords	✅	✅	❌	❌	❌
Assign roles & products	✅	✅	❌	❌	❌
Manage own tenant users	✅	✅	❌	✅	❌
Approve self-registrations	✅	✅	❌	❌	❌
Tenant Management
View all tenants	✅	✅	❌	❌	❌
Create / Edit tenants	✅	✅	❌	❌	❌
Set token quotas	✅	✅	❌	❌	❌
Suspend / Activate tenant	✅	✅	❌	❌	❌
Product Management
View all products	✅	✅	👁 assigned only	👁	❌
Create / Edit products	✅	✅	❌	❌	❌
Chunking configuration	✅	✅	❌	❌	❌
Kill / Activate product	✅	✅	❌	❌	❌
Document Management
Upload documents	✅	✅	✅ assigned only	❌	❌
Reindex / Archive docs	✅	✅	✅ assigned only	❌	❌
View documents	✅	✅	✅ assigned only	👁	❌
API Keys
Generate API keys	✅	✅	❌	❌	❌
Rotate / Revoke keys	✅	✅	❌	❌	❌
View API keys	✅	✅	❌	❌	❌
FAQ Overrides
Create / Edit FAQ	✅	✅	✅ assigned only	❌	❌
View FAQ overrides	✅	✅	✅ assigned only	❌	❌
Analytics & Reports
Platform-wide analytics	✅	✅	👁 assigned only	❌	❌
Token usage by tenant	✅	✅	❌	❌	❌
Token usage by product	✅	✅	✅	❌	❌
Own tenant analytics	✅	✅	👁	✅	❌
Export CSV reports	✅	✅	❌	❌	❌
Audit Logs
View audit logs	✅	✅	❌	❌	❌
Export audit logs	✅	❌	❌	❌	❌
System Settings
LLM provider & model	✅	❌	❌	❌	❌
SMTP configuration	✅	❌	❌	❌	❌
Role management	✅	❌	❌	❌	❌
Appearance settings	✅	✅	✅	✅	❌
Chat / Widget
Admin test chat	✅	✅	✅ assigned only	❌	❌
Super admin chat review	✅	❌	❌	❌	❌
End-user chat (widget)	✅	✅	✅ assigned only	✅ assigned only	✅

Role Descriptions

Super Admin

Complete platform control — LLM settings, kill switch, all tenants, system config. Assigned to NetScore IT only.

Core Admin

Full operational access — manage users, tenants, products, documents. Cannot change LLM or system config.

Product Admin

Manages assigned products only — upload documents, configure FAQ, view product analytics.

Tenant Admin

Manages users within their own tenant. Can view their tenant's usage and documents.

Name	Prefix	Tenant	Type	Rate Limit	Expires	Status	Actions

No API keys found

Filename	Product	Version	Chunks	Size	Status	Uploaded By	Actions
		v
No documents found

Total Tokens

Total Queries

Avg Response

	Tokens Used	Queries	Quota	Usage %	Last Active
				—
No data for selected period

Tenant	Queries	Tokens Used	Quota	Usage %	Avg Response

No data

Product	Queries	Tokens Used	Avg Response	Chunks Found

No data

User	Email	Queries	Tokens Used	Last Active

No data

🔔 System Alerts

Alert types: quotaToken quota approaching limit llm_errorLLM provider error or offline failoverAutomatic provider failover occurred index_failedDocument indexing failed systemGeneral system events

Open Alerts

Type	Severity	Title	Message	Tenant	Time	Action

✅ No open alerts — system is healthy

⚙️ Alert Rules Configuration

📊 Token Quota Alert

Alert when a tenant's token usage exceeds a % of their monthly quota.

Trigger at % of quota

⚡ Configured per-tenant in Tenants → Edit Tenant → Soft Alert %

🤖 LLM Error Alert

Alert when LLM provider fails or becomes unreachable. Automatic failover also triggers an alert.

✅ Always active — cannot be disabled

📄 Document Index Failed

Alert when a document fails to process or embed. Check the Documents page for details.

✅ Always active — cannot be disabled

👤 New User Registration

Alert when a new user self-registers and is awaiting admin approval.

Email notification to admin

📧 Email Notification Settings

Alert emails are sent via your configured SMTP server. Configure SMTP in System → SMTP Settings.

Quota warnings LLM failover events Index failures New registrations

Admin email:

📋 Audit Logs Immutable

User	Role	Action	Target	IP	Time

📧 SMTP Configuration

SMTP Host

Port

Username

Password

From Email

From Name

Test Email (send to)

⚙️ System Controls

Global AI Kill Switch

Immediately blocks all AI chat responses across all tenants.

LLM Health Check

Messages

Tokens used

Avg response

No sessions found

Select a product and start chatting with Medha

(Test mode — non-billable)

Select a session from the left to review

View all user conversations, feedback and responses

No messages in this session

💬 FAQ Overrides

Pinned answers returned instantly without calling the LLM.

Filter by Product:

Question / Trigger	Answer	Product	Match Mode	Status	Actions

No FAQ overrides yet.

🤖 LLM Settings

Configure the language model. Changes apply immediately to all new chat requests.

Provider

🖥️ Ollama Local GPU / CPU

⚡ vLLM High-throughput local

🌐 OpenAI GPT-4o / cloud

🖥️ Ollama Configuration

Ollama URL

Model

Current active:

Keep Alive

Temperature

Max Tokens (num_predict)

Context Window (num_ctx)

⚡ vLLM Configuration

vLLM Server URL

API Key (optional — leave blank if no auth)

Model

Current active:

Temperature

Max Tokens

⚡ vLLM uses the OpenAI-compatible /v1/chat/completions endpoint. Your server: https://vllm.netscoretech.com — Model: qwen2.5-7b-instruct. No API key required.

🌐 OpenAI Configuration

API Key

Model

Temperature

Max Tokens

Base URL (Azure / proxy override)

🧪 Connection Test

Testing…

🎨 Color Theme

Choose your preferred visual theme. Saved automatically and applied on every visit.

Active theme:

🔤 Font Size

💬 Chat Interface

Show token count per message Show source chunks count Show response time

↩️ Reset

Reset all appearance settings to their defaults.

⚙️ Chunking settings apply to all future document indexing. Re-index existing documents after changing these values.

⚙️ Document Chunking Configuration

Product	Collection	Chunk Size (words)	Overlap (words)	Top-K	Similarity Threshold	Actions

📄 Re-Index Documents by Product

Select Product to Re-Index All Documents

🧠 LLM Infrastructure Health

⏳ Pending Approval Requests

🔐 Role Access Matrix

🔔 System Alerts

Open Alerts

⚙️ Alert Rules Configuration

📋 Audit Logs Immutable

📧 SMTP Configuration

⚙️ System Controls

💬 FAQ Overrides

🔍 Chat Query Review

🤖 LLM Settings

Provider

🖥️ Ollama Configuration

⚡ vLLM Configuration

🌐 OpenAI Configuration

🧪 Connection Test

🎨 Color Theme

🔤 Font Size

💬 Chat Interface

↩️ Reset

⚙️ Document Chunking Configuration

📄 Re-Index Documents by Product

✏️ Edit Tenant

🏢 New Tenant

👤 New User

✏️ Edit User

📧 Invite User

🔑 Reset Password

⚙️ Product Settings —

💬 New FAQ Override

✏️ Edit FAQ Override

⬆️ Import FAQ Overrides

📦 New Product

🔐 Role Permissions —

🔐 New Role

🔑 Generate API Key

⬆ Upload Document

🔑 New Key Generated