Skip to main content

Agent Basics

Learn how to create and manage agents in Gentic. This guide covers the fundamentals of agent creation, configuration, and management.

Creating Agents​

Agents are the core building blocks of Gentic. They can be configured with specific capabilities and behaviors.

Basic Agent Creation​

const agent = await client.createAgent({
  name: 'Customer Support Agent',
  description: 'An agent that helps with customer support queries',
  capabilities: ['chat', 'ticket-management', 'knowledge-base']
});

Agent Capabilities​

Agents can have various capabilities:

  • chat: Basic conversational abilities
  • document-search: Search through connected knowledge sources
  • ticket-management: Handle support tickets
  • workflow-execution: Execute defined workflows
  • data-analysis: Analyze and process data
  • api-integration: Connect with external APIs

Configuring Agents​

Setting Agent Parameters​

await client.updateAgent(agent.id, {
  parameters: {
    temperature: 0.7,
    maxTokens: 1000,
    model: 'gpt-4',
    responseStyle: 'professional'
  }
});

Adding System Prompts​

await client.updateAgent(agent.id, {
  systemPrompt: `You are a helpful customer support agent. 
  Always be polite and professional.
  Use the knowledge base to provide accurate information.
  If you're unsure, ask for clarification.`
});

Managing Agent Knowledge​

Connecting Knowledge Sources​

// Connect a knowledge base
await client.connectKnowledgeSource(agent.id, knowledgeBaseId);

// Connect a file collection
await client.connectKnowledgeSource(agent.id, fileCollectionId);

// Connect a knowledge graph
await client.connectKnowledgeSource(agent.id, knowledgeGraphId);

Managing Knowledge Access​

// Set knowledge source priority
await client.updateKnowledgeSourcePriority(agent.id, {
  [knowledgeBaseId]: 1,
  [fileCollectionId]: 2
});

// Configure knowledge retrieval settings
await client.updateAgent(agent.id, {
  knowledgeSettings: {
    maxResults: 5,
    relevanceThreshold: 0.7,
    useReranking: true
  }
});

Workflow Integration​

Assigning Workflows​

// Assign a workflow to an agent
await client.assignWorkflow(agent.id, workflowId);

// Configure workflow triggers
await client.updateAgent(agent.id, {
  workflowTriggers: {
    [workflowId]: {
      trigger: 'on_message',
      conditions: {
        containsKeywords: ['support', 'help', 'issue']
      }
    }
  }
});

Monitoring and Analytics​

Viewing Agent Performance​

// Get agent metrics
const metrics = await client.getAgentMetrics(agent.id);

// Get conversation history
const conversations = await client.getAgentConversations(agent.id, {
  limit: 10,
  offset: 0
});

// Get workflow execution history
const executions = await client.getWorkflowExecutions(agent.id, {
  workflowId: workflowId,
  limit: 10
});

Best Practices​

  1. Start Simple: Begin with basic capabilities and add complexity gradually
  2. Monitor Performance: Regularly check agent metrics and user feedback
  3. Iterate and Improve: Update agent configuration based on usage patterns
  4. Test Thoroughly: Test agents with various scenarios before deployment
  5. Maintain Knowledge: Keep connected knowledge sources up to date

Next Steps​