Skip to main content
Noorle is designed for diverse use cases across AI automation, integrations, and multi-agent systems. Here are some real-world applications:

​
1. AI-Powered Customer Support Agents

Deploy AI agents to answer customer questions while integrating with your internal systems. Architecture:
  • Agent accesses Files capability to read knowledge base documents
  • Agent uses Knowledge Retrieval to semantically search product documentation
  • Agent calls custom plugin to query customer database for order history
  • Agent uses HTTP Client connector to look up real-time inventory
Result: Fully autonomous support with human handoff when needed.
Use Noorle’s three-tier memory to maintain conversation context and summarize long interactions for efficient LLM token usage.

​
2. Multi-Agent Workflows

Coordinate multiple specialized agents working together on complex tasks. Scenario: Research report generation
  • Agent 1 (Researcher) - Uses Web Search to gather information
  • Agent 2 (Analyst) - Calls Code Runner to process data and generate visualizations
  • Agent 3 (Writer) - Assembles findings into polished report
How it works:
  • Agents communicate via A2A protocol through agent gateways
  • Each agent has its own MCP gateway exposing specific capabilities
  • Central coordinator agent routes tasks and manages state
  • Workflow is defined as a durable graph with state machines
Result: Complex multi-step automation that’s resilient to failures and can be paused/resumed.
Coordinate agents using the A2A protocol for reliable inter-agent communication. Always define clear handoff points and state machines for complex workflows.

​
3. REST API Aggregation

Connect disparate REST APIs and expose them through a unified MCP interface. Use Case: E-commerce data platform
  • Connector 1: Shopify API (inventory, orders)
  • Connector 2: Stripe API (payments, subscriptions)
  • Connector 3: SendGrid API (email notifications)
  • Connector 4: Analytics system (custom REST API)
Benefits:
  • AI agents query all APIs without needing to learn each one’s schema
  • Unified authentication and error handling
  • Type-safe tool discovery via MCP
  • Omni Tool handles complex queries like “find unpaid orders and send them an email reminder”
Result: Agents can orchestrate cross-system operations seamlessly.

​
4. Custom Business Logic via Plugins

Build domain-specific tools as WebAssembly plugins. Example: Loan approval engine
  • Plugin analyzes credit scores, income, and debt ratio
  • Implements proprietary business rules
  • Runs in secure WASM sandbox with controlled execution
  • Exposed as tool through MCP gateway
Why WASM?
  • Secure: Sandboxed execution prevents plugin from accessing host system
  • Fast: Near-native performance for complex calculations
  • Portable: Write once, deploy anywhere
  • Versioned: Multiple versions can coexist
Result: Enterprise-grade custom capabilities without compromising security.
Plugins are versioned and can be rolled back. Define clear interfaces via MCP tools so agents know how to use your custom logic.

​
5. ChatOps Integration

Expose agents through messaging platforms (Slack, Teams, Discord). How it works:
  1. User sends message in Slack
  2. Slack bot forwards to Noorle agent via A2A protocol
  3. Agent processes request using all available tools:
    • Files for documentation lookup
    • Code Runner to execute scripts
    • Connectors to check deployment status
    • Custom plugin for internal policies
  4. Agent responds in Slack with results
Result: Employees can interact with AI agents naturally within their existing tools.

​
6. Knowledge-Intensive AI Systems

Build RAG (Retrieval-Augmented Generation) systems with semantic search. Components:
  • Knowledge Base: Vector embeddings of product docs, policies, runbooks
  • Agent: Uses Knowledge Retrieval capability to find relevant documents
  • Memory: Three-tier system keeps recent context while compressing history
  • LLM: Answers questions grounded in retrieved documents
Features:
  • Semantic search finds relevant docs even with fuzzy queries
  • LLM reranker improves result quality
  • Document chunking optimizes embedding quality
  • Memory system prevents LLM from losing context on long conversations
Result: Accurate, context-aware responses with citations.

​
7. Data Pipeline Orchestration

Automate data processing workflows with agents. Pipeline: 
Data Source → Extract (Files) → Transform (Code Runner)
              ↓
         API Connector → Enrich (Custom Plugin)
              ↓
        Write Results (Files/Connector) → Notify (HTTP Client)
Agent responsibilities:
  • Read from S3 via Files capability
  • Transform data using Code Runner (Python/Node)
  • Call external enrichment APIs via connectors
  • Monitor pipeline status and alert on failures
Result: Flexible, maintainable data pipelines that adapt to changes.

​
8. Interactive Development Tool

Agents that help developers build, test, and deploy code. Agent capabilities:
  • Computer: Execute commands, view screen, interact with dev tools
  • Code Runner: Execute and test code snippets
  • Browser: Automate testing of web applications
  • Files: Access project files and write code
  • HTTP Client: Make API calls during development
Use cases:
  • Generate code based on specifications
  • Automatically fix failing tests
  • Deploy applications with confidence
  • Create documentation from code comments
Result: Development augmented by AI for higher productivity.
The Computer capability is agent-only (not exposed via MCP) for security. Only trust agents with direct system access.

​
9. Compliance & Audit Automation

Build agents for regulatory compliance tasks. Capabilities:
  • Knowledge bases with regulations, policies, procedures
  • File system access to audit logs and configurations
  • Custom plugins implementing compliance checks
  • HTTP connectors to compliance tools
Examples:
  • Automated security audit: scan systems, check configs, generate compliance report
  • HIPAA audit trail: monitor access logs, detect anomalies, alert administrators
  • Financial reconciliation: compare transactions across systems, identify discrepancies
Result: Continuous compliance monitoring with audit trails.

​
10. AI-Powered Analytics

Agents that explore data and generate insights. Architecture:
  • Agent accesses knowledge base with business context
  • Code Runner executes data analysis (SQL, Python, R)
  • Agent calls visualization APIs via connectors
  • Results stored in Files for sharing
  • Browser capability to interact with BI tools
Workflow:
  1. User asks question: “Which products have declining sales?”
  2. Agent retrieves relevant metrics from knowledge base
  3. Agent runs SQL analysis via Code Runner
  4. Agent generates visualization via connector
  5. Agent delivers findings with context and recommendations
Result: Self-service analytics that require no manual querying.

​
Choosing Your Use Case

Agent + Capabilities

Start here: Deploy an agent with builtin capabilities (Web Search, Files, Code Runner). Perfect for prototypes.

Connectors for APIs

Integrate REST APIs and MCP servers without writing code. Use Omni Tool for smart discovery.

Custom Plugins

Develop proprietary tools in WASM. For complex business logic, security, or performance.

Multi-Agent Systems

Coordinate agents via A2A for complex workflows. Define state machines for resilience.
Ready to build? Head to the Use tab for step-by-step guides.