RecallSync Single Prompt Voice Agent

RecallSync Voice Agent simplifies launching AI voice agents by eliminating the need for complex webhooks or n8n workflows. It offers built-in call recording, transcription, and performance tracking, all seamlessly integrated using your own provider keys.

Written By ReCallSyncAI

Last updated 5 months ago

Stop building complicated webhooks and fighting with n8n workflows. RecallSync Voice Agent is the easiest way to launch a production-ready AI voice agent using your existing provider keys (Bring Your Own Keys - BYOK) with built-in call recording, transcription, and success criteria tracking. Everything is fully integrated and syncs automatically.

1. Prerequisites

To set up a voice agent, you need to configure the following in RecallSync:

  1. Voice AI Provider: Activate and configure your Text-to-Speech (TTS) and Speech-to-Text (STT) providers (e.g., ElevenLabs, Cartesia, Deepgram) by providing your API keys in Settings > Integrations > Voice AI Solutions. (0:39)

  2. LLM Provider: Activate and configure your LLM provider (e.g., OpenAI, Groq) by providing your API keys in Settings > Integrations > LLM API Keys. (0:52-1:01)

  3. SIP Trunk: Create a SIP Trunk to manage your voice channel.

    • Navigate to Settings > SIP Trunk. (1:04)

    • Click Create New SIP Trunk, fill in the details (Friendly Name, Trunk Name, Phone Numbers), set SIP Credentials, and click Create SIP Trunk. (1:05-1:40) This connects your phone numbers for voice routing.

2. Agent Configuration

Once the infrastructure is set up, configuring the agent is quick and easy.

Create and Configure the Agent

  1. Go to the Agents tab on the left sidebar.

  2. Click to Create or Edit your Voice Agent.

  3. Set the Agent Details:

    • Agent Type: Integrated (1:54)

    • Channel: VOICE_CALL (1:55)

    • Voice Agent Type: Choose Single Prompt Agent for simple flows or Flow Based for complex conversations. (1:56-2:01)

  4. Save Changes. (2:03)

Agent Settings: Model, Voice, and Tools

Inside the Agent Builder, navigate through the settings tabs to define your agent's behavior and capability.

Model Tab (LLM & Prompt) (2:10)

  • Model: Select the LLM you configured (e.g., GPT-4o Mini) that will power the conversation logic.

  • Prompt: Define the agent's identity, role, and conversation goals.

Voice Tab (TTS/STT) (2:14)

  • Voice Provider: Select the TTS provider (e.g., ElevenLabs).

  • Voice: Choose the specific voice for your agent (e.g., a professional or conversational tone).

Settings Tab (Call Configuration) (2:21)

This is where you define how the call is tracked and recorded:

  • Call Summary Prompt: Define the key points you want the AI to capture and summarize after the call is finished (e.g., "Summary of key points discussed"). (2:24)

  • Success Criteria: Clearly define what makes the call successful. This tells the AI when to label a call as a success or failure, removing manual judgment. (2:26)

    • Example Criteria: A call is a success if the user is interested in learning more about RecallSync or has booked an appointment.

Tools Tab (Actions) (2:29)

  • Connect Tools: Add any necessary tools the agent can use during the conversation, such as a Calendar tool for booking meetings.

3. Automatic Call Data Sync (The Magic ✨)

This is the core benefit of using RecallSync. Every voice call made through the agent automatically generates and syncs all critical data without any extra integration work or n8n flows.

Data Point

Sync Location

Feature

Call Recording

Available in the Call Logs tab.

Full recording is stored and linked to the lead and conversation.

Full Transcript

Available immediately after the call.

Provides a word-for-word record of the entire conversation.

Call Summary

Available on the Call Logs and Conversation views.

AI-generated summary based on your defined prompt.

Success Status

Updates the Call/Lead Status.

Automatically marked Successful or Failed based on your Success Criteria.

Conversation History

Stored in the Conversations tab.

Provides a single, clean thread of the AI's conversation with the lead.

4. Testing the Agent

You can instantly test your production agent directly from the builder.

  1. In the Agent Builder, click the Test Call button. (2:35)

  2. Enter the phone number you wish to call for the test.

  3. Click Call Now. (2:39)

The agent will initiate the call, and all call metadata, including recording, transcript, and summary, will sync automatically upon completion (2:41), verifying your setup is production-ready.