---
updatedAt: 2026-05-26T16:20:11.000Z
agentTools:
  projectIndex: https://docs.d-id.com/llms.txt
---

# Overview

Create an agent that delegates the conversation pipeline to ElevenLabs and renders a D-ID avatar

Create an agent backed by an ElevenLabs Conversational AI agent. D-ID renders the expressive avatar; ElevenLabs handles speech-to-text, the LLM, and text-to-speech. You bring an existing ElevenLabs agent and API key.

## How It Works

```mermaid
flowchart LR
    User(["fa:fa-user User"]) -- "audio" --> Avatar(["fa:fa-circle-user D-ID Expressive Avatar"])
    Avatar -- "audio" --> ElevenLabs(["fa:fa-microphone ElevenLabs Agent<br/>STT · LLM · TTS"])
    ElevenLabs -- "speech" --> Avatar
    Avatar -- "video + audio" --> User
```

User audio is forwarded to your existing ElevenLabs agent for the entire conversation pipeline. D-ID receives the synthesized speech back and renders it through the expressive avatar.

## What Each Side Handles

| ElevenLabs                            | D-ID                        |
| ------------------------------------- | --------------------------- |
| Speech-to-text                        | Expressive avatar rendering |
| LLM and prompting                     | Session orchestration       |
| Text-to-speech                        | Embed and Client SDK        |
| Knowledge, tools, conversation config | Vision input (optional)     |

## Connect

Once the agent is created you can connect through two paths:

<Cards>
  <Card title="Embed Code" icon="fa-code">
    Drop a script tag into your page for the prebuilt agent UI with no frontend code.
  </Card>

  <Card title="Client SDK" icon="fa-rocket">
    Install the SDK to build a custom UI with full control over layout and behavior.
  </Card>
</Cards>