How to Enable Custom Agent

This guide provides detailed instructions on configuring and deploying your custom AI agent.

Creating AI Agent Project

Start by navigating to the AppBuilder Console. Here, select a project template designated for Custom AI Agents.

Enable Conversational AI

To configure your project for Conversational AI, follow these steps:

1. Navigate to the Agora Console.
2. Select your project from the list of projects.
3. Click on the Edit button for your project.
4. In the "Conversational AI" section, click on the "Enable Conversational AI" toggle button.

After activation, you can add the agent in your project by following the steps below:

Configure AI Agent Project

1. Custmoize Agent: Edit the default agent or create a new one .

2. Customize System Prompt: This prompt is a to define the agent's behavior and provide context to your custom AI agent.

3. Edit Greeting Prompt: The greeting prompt is a message that appears at the start of an interaction with your custom AI agent.

4. Edit Failure Prompt: This prompt is a message that appears when an interaction with your custom AI agent fails.

5. Select ASR Language: Choose the language you want to use for ASR (Automatic Speech Recognition).

6. Intelligent Interruption Handling: This feature allows you to manage interruptions more effectively by enabling graceful interruption and suppressing background human voices.

7. LLM Settings: This feature allows you to configure settings for your LLM (Language Model)

   • Select a provider from the settings
   • Select a model from the settings
   • Enter a URL for the LLM
   • Enter the API key for the LLM If you want to use your own LLM , then select Others from Providers and enter the URL for your LLM.It should be compliant with the OpenAI Protocol.

8. TTS Settings: This feature allows you to configure settings for your TTS (Text-to-Speech). Currently, you can choose between ElevenLabs and Microsoft Text-to-Speech.

9. VAD Settings: The available VAD (Voice Activity Detection) settings include parameters for managing human voice interruptions.

10. Custom Attributes: Any additional custom attributes can be added here.

Deploy AI Agent Project

Once your project is configured, deploy it using the deployment option available in the top bar menu.

Managing Data in Your Custom Agent

Here are some example how you can build your own custom agent and intergarte that with your appbuilder project. You will need to deploy your custom agent and provide the deployed URL in the LLM settings of your appbuilder project.

Example : Restuarant Customer Suppport Agent

The agent is designed to provide customer support for a restaurant.It can handle a wide range of customer inquiries and provide assistance. It integrates seamlessly with app builder projects, includes RAG App capabilities, where PDFs can serve as data sources for information context. Explore its features and implementation details on GitHub: AI_BUILDER_AGENT.

Example : Meeting Summarizer Agent

The Meeting Summarizer Agent is designed to condense meeting discussions into concise summaries and provide actionable insights, including action items or minutes of the meeting. Explore its features and implementation details on GitHub: AI_BUILDER_AGENT.