Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.subverseai.com/llms.txt

Use this file to discover all available pages before exploring further.

Overview

Subverse supports four agent types, each designed for a different channel and use case:
Agent TypeChannelBest For
Voice AgentPhone callsInbound/outbound calling, voice-based support
Chat AgentWhatsApp / Web chat / Telegram / SMSText-based support, FAQ bots, lead qualification
Email AgentEmail inboxAuto-responding to incoming emails
Background AgentLLM Functions / Workflows / APIAsync task automation, sub-agents in Agentverse
To create any agent, navigate to Agents in the left sidebar and click Add Agent. You’ll be prompted to select the agent type first.
Not sure which type to pick? If you’re just getting started, Voice Agent is the most fully-featured option and a great way to learn the platform. For a text-first use case, start with Chat Agent.

Quick Start: Your First Voice Agent

This guide walks you through creating and testing a Voice Agent — the fastest way to see Subverse in action.

Step 1: Create the Agent

  1. Go to Agents in the left sidebar.
  2. Click Add Agent and select Voice as the agent type.
  3. Give your agent a Name (lowercase letters, numbers, and underscores — e.g. support_bot).
  4. Optionally add a Description for internal reference.

Step 2: Write the System Prompt

The system prompt is the most important setting. It defines your agent’s identity, role, and behavior. Navigate to the Instructions section and fill in:
  • Initial Message — what the agent says when a call connects (e.g. "Hello! You've reached Acme support. How can I help you today?")
  • System Context — the full behavior prompt for your agent. Click on “Generate using AI” and describe the agent behavior and it will generate a great prompt for you.
A good system prompt covers:
  1. Identity — who is the agent?
  2. Role — what is its purpose?
  3. Tone — how should it communicate?
  4. Capabilities — what can it do?
  5. Limits — what should it avoid or escalate?

Step 3: Configure the LLM

Under LLM Service, select a provider (e.g. OpenAI) and a model (e.g. gpt-4o). The default temperature of 0.5 works well for most support use cases.

Step 4: Configure Speech (STT & TTS)

Voice agents require:
  • STT Service — how the agent hears the caller (e.g. Deepgram, en-US)
  • TTS Service — how the agent speaks (e.g. ElevenLabs, choose a Voice ID)
For a quick start, pick any provider, select a language, and choose a voice. You can fine-tune these later.

Step 5: Save and Test

  1. Click Save in the top-right corner.
  2. To test, open the Call Sidebar on the agent’s page.
  3. Enter the bot’s phone number and your personal number, then click Call.
  4. The agent will call your phone — speak naturally and verify the behavior.

Next Steps

Once your first agent is running, explore the full configuration options or set up another agent type:

Voice Agent

Full guide: STT, TTS, turn management, call transfers

Chat Agent

Text-based agents for web chat and messaging

Email Agent

Auto-reply agents for incoming email

Background Agent

Async task agents for workflows and automation