Skip to main content
A PAL (Personified Application Layer) contains the full configuration for its role in CVI: behavior, knowledge, objectives, guardrails, tools, and every pipeline layer. The face defines the on-screen likeness and voice; the PAL defines how it thinks, speaks, and acts.
Terminology update: Tavus now uses PAL (behavior, knowledge, and pipeline configuration) and Face (visual appearance and voice) in the API and docs. Legacy names persona and replica still work on existing endpoints and request fields (/v2/personas, /v2/replicas, persona_id, replica_id, and related aliases) for backward compatibility.
At a glance
  • PAL - The agent’s identity (name, behavior) plus pipeline mode, default face, layers, documents (Knowledge Base), objectives, and guardrails. Voice comes from the face by default; the TTS layer can override it.
  • Relationship to CVI - PALs hold the settings that drive a real-time CVI session; see What is CVI? for the full stack (WebRTC, layers, Phoenix, and latency characteristics on the default path).
  • Layers (order of guides below) - Perception → STT → Conversational Flow → LLM → TTS; each has its own configuration page.

​
PAL Customization Options

Each PAL includes configurable fields. Here’s what you can customize:
  • PAL Name: Display name shown when the PAL joins a call.
  • System Prompt: Instructions sent to the language model to shape the PAL’s tone, personality, and behavior.
  • Pipeline mode: Controls which CVI pipeline layers are active and how input/output flows through the system. See Pipeline modes for how the full pipeline, Echo, integrations, and custom LLM paths differ.
  • Default face: Sets the photorealistic face associated with the PAL (default_face_id on POST /v2/pals; required).
  • Layers: Perception, STT, conversational flow, LLM, and TTS - each processes part of the interaction and can be tuned independently (see Layers below).
  • Documents: A set of documents that the PAL has access to via the Knowledge Base (retrieval-augmented generation, RAG).
  • Objectives: The goal-oriented instructions your PAL will adhere to throughout the conversation.
  • Guardrails: Conversational boundaries that can be used to strictly enforce desired behavior.

​
Objectives & Guardrails

Provide your PAL with robust workflow management tools, curated to your use case

Objectives

The sequence of goals your PAL will work to achieve throughout the conversation - for example, gathering a piece of information from the user.

Guardrails

Conversational boundaries that can be used to strictly enforce desired behavior.

​
Layers

Explore our in-depth guides to customize each layer to fit your specific use case:

Perception Layer

Defines how the PAL interprets visual input like facial expressions and gestures.

STT Layer

Transcribes user speech into text using the configured speech-to-text engine.

Conversational Flow Layer

Controls turn-taking, interruption handling, and active listening behavior for natural conversations.

LLM Layer

Generates PAL responses using a language model. Supports Tavus-hosted or custom LLMs.

TTS Layer

Converts text responses into speech using Tavus or supported third-party TTS engines.

Conferencing Layer

Gives the PAL an email identity so it can be invited to Google Meet calls via Google Calendar and join automatically - including meetings already in progress.