Operators are persistent chatbot-like assistants within a workspace. Each combines a specific purpose with skills, tools, MCP servers, and knowledge bases to function as a virtual team member.
Creating an Operator
- Launch StationOne and select a workspace.
- Right-click the workspace → Settings → Operators.
- Click + to create a new operator.
- Fill in the fields:
- Name (required): A unique name for the operator.
- Description (required): What the operator does.
- AI Model: Select the LLM engine and model.
- Instructions: Define the operator’s behavior covering:
- Purpose/Objectives: Primary purpose and key tasks.
- Interaction Design: Conversational flow, tone, response handling.
- Performance/Governance: KPIs, metrics, compliance constraints.
- Skill: Attach a skill matching the operator’s purpose.
- Tools: Select which tools (plugins and MCP server tools) the operator can use.
- Knowledge Base: Attach one or more knowledge base collections for RAG.
- Click Save.
Editing an Operator
- Right-click workspace → Settings → Operators.
- Click the edit icon on the target operator.
- Update any field: Name, Description, AI Model, Instructions, Skill, Tools, Knowledge Bases.
- Click Save.
Using an Operator
- Click Operators in the workspace sidebar.
- Select the operator from the list.
- Chat with the operator as you would in a regular conversation.
- The operator uses its configured skill, tools, and knowledge bases automatically.
Multiple Knowledge Bases
Operators support attaching multiple knowledge base collections. When a user asks a question, the operator searches all attached knowledge bases and combines results for a comprehensive answer.
MCP Server Access
Operators have access to MCP server tools configured in the workspace. When creating or editing an operator, the Tools selection includes both native plugins and connected MCP server tools.