Imported from jeremylongshore/tons-of-skills-marketplace (plugins/saas-packs/retellai-pack/skills/retellai-core-workflow-a/SKILL.md). Install upstream with npx skills add jeremylongshore/tons-of-skills-marketplace --skill retellai-core-workflow-a. Copyright stays with the author (MIT).

Retell AI Core Workflow A

Overview

Build and configure voice agents with custom prompts, function calling, and call flow logic.

Prerequisites

• Completed retellai-hello-world

Instructions

Step 1: Agent with Function Calling

const llm = await retell.llm.create({
  model: 'gpt-4o',
  general_prompt: `You are a booking assistant for Dr. Smith's office.
    - Help callers book, reschedule, or cancel appointments
    - Collect: name, phone, preferred date/time
    - Confirm all details before booking`,
  functions: [
    {
      name: 'book_appointment',
      description: 'Book a new appointment',
      parameters: {
        type: 'object',
        properties: {
          patient_name: { type: 'string' },
          phone: { type: 'string' },
          date: { type: 'string', description: 'YYYY-MM-DD format' },
          time: { type: 'string', description: 'HH:MM format' },
        },
        required: ['patient_name', 'phone', 'date', 'time'],
      },
      url: 'https://your-api.com/appointments',
      speak_during_execution: true,
      speak_after_execution: true,
    },
  ],
});

Step 2: Configure Voice and Behavior

const agent = await retell.agent.create({
  response_engine: { type: 'retell-llm', llm_id: llm.llm_id },
  voice_id: '11labs-Rachel',
  agent_name: 'Dr. Smith Booking Agent',
  language: 'en-US',
  opt_out_sensitive_data_storage: false,
  end_call_after_silence_ms: 10000,  // End call after 10s silence
  max_call_duration_ms: 300000,       // 5-minute max
  enable_backchannel: true,           // "mhm", "yeah" responses
  boosted_keywords: ['appointment', 'schedule', 'Dr. Smith'],
});

Step 3: Update Agent Configuration

await retell.agent.update(agent.agent_id, {
  voice_id: '11labs-Dorothy',  // Change voice
  end_call_after_silence_ms: 15000,
});

Output

• Agent with custom LLM prompt and function calling
• Voice and behavior configuration
• Real-time function execution during calls

Error Handling

Examples

Create and validate a new appointment-intake agent

Build the agent in a development project with a narrow appointment-booking tool and a synthetic calendar. Run representative happy-path, no-availability, and human-handoff calls, checking that each tool request uses the expected arguments. Review the redacted transcripts with the service owner, then create a versioned preview for a limited internal canary; keep the previous agent version available until the new flow meets its completion and transfer goals.

Resources

• Retell AI Documentation
• retell-sdk npm

Next Steps

Phone call management: retellai-core-workflow-b