​
What is AMS?

The Agent Management Studio (AMS) is a visual, no-code admin portal for building, testing, and monitoring AI agents. You can use AMS to control workflow access, manage MCP servers, upload data and guardrails, shape personality and behavior, and deploy agents quickly without writing code.

​
AMS Key Components at a Glance

​
Agents

Autonomous units tailored to a role or objective. Agents use assigned Workflows, MCP Servers, and Knowledge Contexts to act like subject matter experts. Learn more about agents here When creating an agent, you need to configure all attributes.
More information can be found here

​
Knowledge Contexts

Bundles of files, links, and data that an agent can retrieve at runtime. This is the agent’s dynamic reference library. Find more information about Knowledge Contexts here

​
MCP Servers

Model Context Protocol (MCP) servers standardize how agents talk to tools, APIs, and data sources. They bundle tools and expose Workflows to agents in a consistent way. Learn more about MCP Servers here

​
Workflows

Workflows transform how you build agent capabilities. Instead of complex development, you describe what you want your agent to do in natural language and the system generates functional tools automatically. More about Workflows here

​
End-to-End: Build and Deploy an Agent

​
Step 1: Create a Workflow (Workflow Builder)

  • Go to Workflows → New Workflow
  • Describe what the workflow should do in plain language
  • Refine until the preview looks correct

​
Step 2: Add the Workflow to an MCP Server

  • Open MCP Servers and create or select a server
  • In Workflows, attach the Workflow to the MCP server
  • Save to expose the workflow via this MCP server

​
Step 3: Create your Agent

  • Go to Agents → New Agent
  • Set name, role, or objective, and personality or guardrails
  • Save

​
Step 4: Create a Knowledge Context

  • Go to Knowledge Contexts → New Context
  • Name and describe your new context
  • Upload content into the knowledge context

​
Step 5: Attach the Knowledge Context to the Agent

  • Open your Agent → Contexts tab
  • Add one or more Knowledge Contexts

​
Step 6: Attach the MCP Server to the Agent

  • Go to Agent → MCP Servers tab
  • Add one or more MCP servers to the agent

​
Step 7: Deploy the Agent

  • Go back to the dashboard to see your live agents
  • Select the agent you want to deploy
  • Test it out if needed
Then, go to the Deployment tab to complete deployment. Learn more about deploying your agents here

​
Notes

To view agents on the demo page:
  • Anonymous access must be enabled
  • The agent must be set to Live