KingCaller

Agent Builder

Configure LLM, voice, VAD, and call handling settings for your AI agents.

Agent Builder

The Agent Builder is the comprehensive configuration panel for your voice AI agents. Access it by clicking on any agent from the A.I Agents list.

LLM Settings

Configure the AI model powering your agent's responses:

app.kingcaller.ai
LLM Settings panel showing Model dropdown (GPT-4o selected), Temperature slider at 0.7, Max Tokens input set to 300, and a large System Prompt textarea with custom instructions
Configure the AI model, temperature, and system prompt
SettingDescriptionTypical Values
ModelWhich AI model to useGPT-4o, GPT-4o Mini, Claude
TemperatureCreativity vs consistency0.1 (precise) – 1.0 (creative)
Max TokensMaximum response length150–500 tokens
System PromptAgent personality and instructionsYour custom instructions

Speech Settings (TTS/STT)

Text-to-Speech (TTS)

Configure how your agent speaks:

app.kingcaller.ai
TTS Settings showing Provider dropdown (ElevenLabs), Voice selector with preview button and waveform, Speed slider (0.5x to 2.0x), Stability slider, and Language selector
Configure voice, speed, and stability for Text-to-Speech
SettingDescription
TTS ProviderElevenLabs, OpenAI, or others
VoiceSelect from the Voice Library
SpeedSpeaking rate (0.5x – 2.0x)
StabilityVoice consistency vs natural variation
LanguagePrimary language of the agent

Speech-to-Text (STT)

Configure how your agent listens:

SettingDescription
STT ProviderDeepgram, OpenAI Whisper, etc.
LanguageLanguage to transcribe
ModelAccuracy vs speed tradeoff

VAD Settings (Voice Activity Detection)

Control how the agent detects when the caller is speaking:

SettingDescriptionDefault
SensitivityHow easily silence triggers a responseMedium
Silence DurationHow long to wait before responding (ms)500ms
Prefix PaddingAudio captured before speech detection300ms

Call Settings

app.kingcaller.ai
Call Settings panel with toggle switches for Call Recording (ON), Voicemail Detection, DTMF, Call Transfer, Noise Suppression, plus Max Duration input and Interruption Mode dropdown
Configure call recording, transfer, DTMF, and more
SettingDescription
Max DurationAuto-end call after time limit (e.g., 15 min)
Interruption ModeHow agent reacts when caller interrupts
Call RecordingRecord calls for playback and analysis
Voicemail DetectionDetect answering machines on outbound calls
DTMFHandle keypad input (press 1 for sales, etc.)
Call TransferTransfer to human agents or external numbers
Noise SuppressionFilter background noise

Saving Changes

After configuring settings:

  1. Click Save Changes at the bottom
  2. If the agent is active, changes take effect on the next call
  3. Test your changes with a test call

On this page