Enterprise-Grade AI Agent Development Kit

Build fault-tolerant AI agent workflows with temporal.io reliability. Platform-agnostic architecture with MCP tool integration and multi-agent coordination.

> dotnet add package XiansAi.Lib
Agent Visualization

Agent Visualization

Design and monitor complex AI workflows with intuitive visualization tools. Track agent interactions and data flows in real-time for complete operational visibility.

Agents in a Team

Case study – Sales Prospecting Agent Team

Enterprise AI workflows require specialized agents working in coordination. Built on temporal.io for fault-tolerant execution, agents communicate through MCP (Model Context Protocol) for tool integration andAgent-to-Agent interactions for seamless handoffs between different capabilities.

Agent Composition Architecture

Create Your First Multi-Agent Team

Build enterprise-grade AI workflows with temporal.io reliability. Here's how to create a fault-tolerant sales prospecting agent team:

  • Define specialized agents with MCP tool integration
  • Coordinate Agent-to-Agent communication workflows
  • Execute with temporal.io fault tolerance and durability


      // Create the agent team with temporal.io workflow engine
      var agentTeam = new AgentTeam("SalesProspectingTeam");
      
      // Scheduler agent with fault-tolerant execution
      var schedulerAgent = agentTeam.AddAgent<SchedulerBot>();
      schedulerAgent.SetScheduleProcessor<SchedulerProcessor>(
          processInWorkflow: true, 
          startAutomatically: true, 
          runAtStart: true
      );
      schedulerAgent.AddActivities<ProspectingActivities>();
      
      // Content generation agent
      var contentAgent = agentTeam.AddAgent<ContentFlow>();
      contentAgent.AddActivities<ContentFlowActivities>();
      
      // Interactive assistant with MCP tools
      var assistantAgent = agentTeam.AddAgent<InteractionBot>();
      assistantAgent.AddCapabilities(typeof(SalesCapabilities));
      
      // Web research agent with MCP integration
      var webAgent = agentTeam.AddAgent<WebResearchBot>();
      webAgent.AddKernelModifier(new FirecrawlMCP());
      webAgent.AddKernelModifier(new GoogleSearchMCP());
      webAgent.AddKernelModifier(new PlaywrightMCP());
      
      // Reporting agent with Office 365 integration
      var reporterAgent = agentTeam.AddAgent<ReportingBot>();
      reporterAgent.AddKernelModifier(new MicrosoftO365MCP());
      reporterAgent.AddKernelModifier(new PdfGeneratorMCP());
      
      // Run with enterprise-grade reliability
      await agentTeam.RunAsync();
    

Read documentation to get started

Learn more about the Xians.ai platform and how to use it's capabilities to build your own AI Agents and Workflows.

Read Documentation

Escape Vendor Lock-In with Open Architecture

Deploy AI agents from any source—whether from Xians.ai, other providers, or your own development team. Any containerized agent integrates seamlessly with our platform.

Temporal.io Workflow Engine

Built on temporal.io for enterprise-grade reliability. Workflows survive restarts, handle failures gracefully, and scale horizontally across your infrastructure.


// Temporal.io powered durability - workflows survive restarts
await Workflow.DelayAsync(TimeSpan.FromDays(365));
// Your workflow continues exactly where it left off after a year

MCP Tool Ecosystem Integration

Leverage the growing MCP (Model Context Protocol) ecosystem with seamless tool integration for enterprise workflows.


// MCP tools integrate seamlessly
webAgent.AddKernelModifier(new FirecrawlMCP());
webAgent.AddKernelModifier(new GoogleSearchMCP());
webAgent.AddKernelModifier(new PlaywrightMCP());

reporterAgent.AddKernelModifier(new MicrosoftO365MCP());
reporterAgent.AddKernelModifier(new PdfGeneratorMCP());
// Extensive MCP ecosystem support...

Balance AI Autonomy with Human Control

  • Maintain control over critical business processes
  • Implement AI automation strategically after review
  • Keep human oversight in workflow design
  • Blend autonomy and orchestration seamlessly

Sophisticated Agents, Not Just Chatbots

Our platform is designed to host sophisticated agents capable of executing complex tasks and workflows, far beyond simple chat interactions.


Agent Anatomy

A workflow within an agent intelligently orchestrates between multiple tools that are augmented by prompts

Multi-Workflow AgentAgents can consist of multiple workflows, some for user interaction, others for business process capabilities.

Self-Host in Your Cloud

Deployment ArchitectureAgent Platform Deployment

Deploy the Xians.ai platform and agents in your own infrastructure for complete control and compliance.

Keep sensitive data within your environment
Deploy as containers or app services
Seamless integration with your existing cloud resources
Customize networking and security policies

Built for Multi-Tenancy

Scale your AI operations across multiple customers with built-in tenant isolation. Deploy and manage agents per customer while maintaining unified oversight.

  • Isolate agent execution between tenants
  • Activate or deactivate agents on a per-tenant
  • Maintain separate configurations for tenants
  • Centralized monitoring across tenants

Enterprise Agent Platform Comparison

See how Xians.ai compares to other enterprise AI agent development platforms:

Xians.aiOther Agent Kits
Multi-Agent Coordination
Agent-to-Agent Communication
Temporal.io Workflow Engine?
Durable Workflow Execution?
Long-running Process Support?
MCP Tool Integration
Event-driven Architecture?
Multi-tenancy Support?
Platform Agnostic
Enterprise Fault Tolerance?
Workflow Visualization?
Self-hosted Deployment
Workflow State Persistence?
Management Portal?
Prompt Base Management?
Execution History & Tracing?
Horizontal Scaling?