Enterprise Edition

​

Complete access control system

• Create and manage users with unique credentials
• Assign users to multiple roles
• Control access at the resource level

• Define custom roles with specific permissions
• Grant read/write access to contexts and agents
• Support for private and public resources

• Control who can read/write to knowledge bases
• Set default permissions for new contexts
• Override permissions per user or role

• Multi-tenant SaaS: Isolate customer data
• Enterprise teams: Different departments access different knowledge bases
• Client services: Separate agent access per client
• Compliance: Meet data access audit requirements

// RBAC is configured at the ExuluApp level
await app.create({
  config: {
    authentication: {
      enabled: true,
      jwt: {
        secret: process.env.JWT_SECRET!
      }
    }
  }
});

​

Production-grade background processing

• Embedding generation queues
• Document processing pipelines
• Rate limiting and throttling
• Priority queues
• Automatic retries with backoff
• Dead letter queues
• Job scheduling and cron

• Real-time queue dashboard
• Job status tracking
• Performance metrics
• Worker health monitoring
• Manual job retry/cancel

• Processing thousands of documents
• Batch embedding generation
• Scheduled data updates
• Asynchronous agent tasks
• Resource-intensive operations

await app.create({
  config: {
    workers: {
      enabled: true,
      concurrency: 5,
      telemetry: {
        enabled: true
      }
    }
  }
});

​

Test and measure agent quality

• Create test cases and golden datasets
• Run evaluation sets against agents
• LLM-as-a-judge evaluation
• Custom evaluation functions
• Compare agent versions
• Track metrics over time

• Accuracy and correctness
• Hallucination detection
• Response relevance
• Tool usage correctness
• Latency and performance

• CI/CD integration
• Regression testing
• A/B testing
• Performance benchmarking

import { ExuluEval } from "@exulu/backend";

const evaluation = new ExuluEval({
  id: "customer-support-eval",
  name: "Customer Support Quality",
  testCases: [
    {
      input: "How do I reset my password?",
      expectedOutput: "Contains password reset steps"
    }
  ]
});

​

Reusable multi-step workflows

• Save any conversation as a template
• Parameterized inputs and outputs
• Multi-step reasoning patterns
• Scheduled execution with cron
• Template versioning
• Share across teams

• Standardized report generation
• Recurring analysis tasks
• Multi-step data processing
• Customer onboarding flows
• Automated quality checks

1. Have a conversation with your agent
2. Save it as a template
3. Define input parameters
4. Schedule or trigger on-demand
5. Collect outputs automatically

​

Intelligent multi-step search

• Query preprocessing and expansion
• Multi-strategy search (vector + keyword + hybrid)
• Iterative query refinement
• Result reranking and fusion
• Context-aware filtering
• Source attribution

• Automatic query decomposition
• Ambiguity detection and clarification
• Temporal and entity awareness
• Cross-reference resolution

• Higher accuracy on complex queries
• Better handling of ambiguous questions
• Improved recall for multi-faceted topics
• More relevant context for RAG

// Agentic retrieval is automatically enabled when licensed
// Configure search strategies in your context
const context = new ExuluContext({
  id: "docs",
  embedder: openAIEmbedder,
  reranker: cohereReranker,
  // Agentic retrieval uses multiple search strategies
});

​

Enterprise document parsing

• PDFs (text and image-based)
• Microsoft Office (DOCX, XLSX, PPTX)
• Spreadsheets with table extraction
• HTML with structure preservation
• Markdown with formatting
• Source code with syntax awareness

• Table detection and extraction
• Image and diagram extraction
• Header/footer handling
• Multi-column layouts
• OCR for scanned documents
• Metadata extraction

• Structure-preserving chunking
• Semantic boundary detection
• Reference link preservation
• Source attribution

​

Collect user feedback at scale

• Thumbs up/down on responses
• Written comments and notes
• Rating scales (1-5 stars)
• Custom feedback fields
• Session-level feedback

• Aggregate feedback scores
• Identify problem patterns
• Track improvement over time
• Export feedback data
• Build golden datasets from positive examples

• GraphQL API for feedback
• UI components for feedback collection
• Webhook notifications
• Export to evaluation framework

​

Orchestrate complex agent systems

• Agent-to-agent delegation
• Intelligent routing
• Parallel agent execution
• Result aggregation
• Failure handling and fallbacks

• Request/response
• Publish/subscribe
• Broadcast
• Sequential pipelines
• Parallel fan-out

• Specialized expert agents
• Research and synthesis workflows
• Multi-perspective analysis
• Verification and validation chains
• Complex decision-making systems

​

White-label your deployment

• Custom logos and favicons
• Color scheme customization
• Typography and fonts
• Custom CSS and styling
• Powered-by attribution removal

• Admin UI
• Frontend interface
• Email templates
• PDF exports
• Error pages