Global Settings

Global Settings is where super-admins configure system-wide defaults: plugin instances, resource pricing, rate limits, and BI classification.

note

The Settings page requires Super-Admin role access.

The settings page is organized into tabs.

tip

For model catalog and provider credentials management, see Models.

Plugin Instances Tab

Manage global plugin instances that provide file storage, embeddings, vector search, sandboxing, document processing, and file generation.

Plugin Types

Type	Description
Storage	File storage backends
Embedding	Text embedding models
Vector Store	Vector databases for semantic search
Processor	Document parsing and extraction
Generator	File generation (PDFs, images, etc.)
Sandbox	Sandboxed code execution environments

Filtering

Filter by plugin type using tabs: All, Storage, Embedding, Vector Store, Processor, Generator, Sandbox.

Managing Instances

• Browse Plugin Classes — View available plugin classes and their capabilities
• Create Instance — Create a new instance from a plugin class
• Edit — Modify instance configuration
• Test Health — Verify the instance is working
• Set as Default — Make this instance the system default for its type
• Delete — Remove the instance

Each instance shows its name, description, type, plugin class, status (Active/Default/Inactive), and tags.

Resource Pricing Tab

Configure pricing rules for resource consumption tracking and billing.

Pricing Table

Filter by resource type (All, LLM, Embedding, Plugin) using tabs.

Each configuration shows:

Column	Description
Resource	Resource name and type
Pricing Mechanism	Subscription, Flat, or Tiered
Rates	Summary of pricing rates
Reset Policy	When usage counters reset
Limits	Soft and hard usage limits
Status	Active/Inactive

Pricing Mechanisms

Mechanism	Description
Subscription	Fixed cost per period
Flat	Simple rate per unit
Tiered	Volume-based pricing with multiple tiers

Adding a Configuration

Click Add Configuration to create a pricing rule.

Basic section:

• Resource type, ID, display name, status

Pricing section:

• Pricing mechanism, rates, tiers (for tiered pricing), reset policy

Limits section:

• Soft limit (warning threshold by quantity or cost)
• Hard limit (maximum allowed by quantity or cost)
• Alert webhook URL for limit notifications

Seed Defaults

Click Seed Defaults to auto-populate pricing configurations for all registered models based on their input/output costs.

Rate Limits Tab

Configure system-wide request rate limiting.

Global Configuration

• Enable/Disable toggle for global rate limiting
• Request Limits:
  • Per Second (burst limit)
  • Per Minute (sustained limit)
  • Per Hour (quota limit)
• Upload Limits: Per Minute, Per Hour
• Download Limits: Per Minute, Per Hour

Override Configurations

A table showing assistant-specific or role-specific rate limit overrides.

info

Rate limits follow a resolution order: Role → Assistant → Global → Environment Defaults. The most specific matching limit applies.

Business Intelligence Tab

Configure BI classification and knowledge gap detection.

Classification Settings

Setting	Description
Enable Classification	Toggle automatic conversation classification
Classification Model	LLM used for classification (models grouped by Budget/Standard/Premium with cost displayed)
Classification Schedule	When to run classification (presets: Hourly, Every 6 hours, Daily, Weekly, or custom cron)
Batch Size	Number of sessions to classify per batch

Knowledge Gap Detection

Setting	Description
Enable Gap Detection	Toggle automatic knowledge gap detection
Similarity Threshold	How similar questions must be to group together (slider: 50–99%)
Embedding Model	Model used for semantic similarity matching

Recommended Configuration

• Classification Model: Use a budget model (e.g., Claude Haiku) for cost-effective classification
• Schedule: Daily at 2 AM to avoid peak usage hours
• Batch Size: 30–50 sessions balances throughput and API costs
• Similarity Threshold: 85% is a good starting point; increase if getting too many false positives

Next Steps

• Models — LLM model catalog and provider credentials
• Users & API Keys — Manage portal users
• Usage & Billing — Monitor costs
• Business Intelligence — View BI analytics