Loading navigation...
Unify Agentic AI
Logo
Overview

Overview

Logo

6 minutes READ

Overview

Copilot in UnifyApps acts as a centralized interface, allowing users to access and manage all their agents in a single place. This streamlines workflows by providing a unified environment for interacting with various AI agents, eliminating the need to switch between different interfaces.

Image
Image

Key Capabilities of Copilot

Copilot offers several key functionalities to enhance your interaction with AI agents:

  • Accessing Multiple Agents & Team of Agents in One Place: Copilot consolidates all your agents and team of agents, making them easily accessible from a single dashboard. This eliminates the need to navigate to individual agent pages for interaction.

Image
Image
  • Switching Between Agents: Within Copilot, you can effortlessly switch between different agents. This is particularly useful when you need to consult various specialized agents for different tasks or inquiries.

  • Changing the LLM Model: Copilot provides the flexibility to change the underlying Large Language Model (LLM) used by an agent. This allows you to experiment with different models to optimize agent performance or explore diverse generative capabilities.

Image
Image
  • Managing Conversations:

  • New Chat: You can initiate a new chat session with any selected agent.

  • Chat History: Copilot maintains a history of your conversations with agents, allowing you to review past interactions and track progress.

Image
Image
  • Uploading Attachments: Users can upload attachments directly within the Copilot interface, enabling agents to process and respond to queries that require external file context.

  • Using Tools: Copilot integrates various tools that agents can leverage to perform their tasks. Examples include Web Search and Deep Research capabilities. Users can customize these tools as needed.

Image
Image
  • Voice Input: By clicking on the voice icon, you can speak your prompt which will be translated into text for the agent to process.

Deploying Agents to Copilot

To make an agent available within Copilot, you need to deploy it from the agent's configuration settings. When you create an agent, you'll find an option to publish it, which includes deploying it to Copilot. During the deployment process, you can Tag this version of the agent and provide Version notes to describe changes. Crucially, there will be a checkbox option, Deploy this agent on Copilot, which you must select.

After deployment, agents can be shared with specific users or teams. You can add users or teams to grant them access, and their access level (e.g., Member) can be managed.

Testing Agents

UnifyApps offers two primary ways to test your agents:

  • Testing Individual Agents (Pre-Deployment): Before deploying an agent to Copilot, you can test it individually. This is done by clicking on 'test the agent' within the agent's configuration interface. This allows you to verify its functionality in isolation.

  • Testing Multiple Agents in Copilot: Copilot simplifies the testing of multiple agents simultaneously. Instead of selecting and testing each agent individually, Copilot provides a unified interface where you can easily switch between and interact with various deployed agents to observe their collective performance and interactions.

FAQs

Can I test multiple agents at once in UnifyApps?

Yes, Copilot acts as a unified interface where you can easily switch between and test multiple agents without having to select each one individually.

How do I make my agent available on Copilot?

When creating or publishing an agent, you will see an option to Deploy this agent on Copilot. Selecting this option during the publishing process will make your agent accessible in Copilot.

Can I change the LLM model for an agent within Copilot?

Yes, Copilot allows you to change the underlying LLM model used by an agent, providing flexibility for optimizing performance or exploring different AI capabilities.

Can I use voice to interact with agents in Copilot?

Yes, Copilot supports voice input. By clicking the voice icon, you can speak your prompt, and it will be transcribed into text for the agent.