> ## Documentation Index
> Fetch the complete documentation index at: https://docs.envole.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Multi-Assistant Collaboration Overview

> Introduction to cross-agent collaboration via SSE API

## Prerequisites

This document assumes you have already implemented basic Envole API integration following our [API Integration Guide](https://docs.envole.ai/getting-started-section/api-integration), including:

* Authentication setup with Bearer tokens and X-Envole-User-Id headers
* Thread creation and basic message posting
* EventSource setup for SSE streaming
* Basic event processing (AGENT\_RESPONSE\_CHUNK, AGENT\_RESPONSE\_COMPLETE)

## Overview

Cross-agent collaboration enables multiple specialized agents to work together on complex tasks through a simple @-mention system. When users mention multiple agents in a message (e.g., "@product and @engineering, work together on this feature spec"), the platform automatically orchestrates a collaborative workflow where each mentioned agent contributes their expertise, and optionally provides a unified response.

## Agent Handle System

### What are Agent Handles?

Agent handles are unique identifiers (like "@product", "@support", "@legal") that allow users to invoke specific agents in conversations. Each agent must have a unique handle within an organization.

### Handle Structure

```typescript theme={null}
interface Agent {
    id: string;
    name: string;
    handle: string;  // Unique handle without @ symbol (e.g., "product")
    // ... other properties
}
```

### Using Handles in Messages

Users invoke collaboration by mentioning agents in their messages:

```javascript theme={null}
// Single agent
"@product what features should we prioritize next quarter?"

// Multiple agents (triggers collaboration)
"@product and @engineering, can you work together on the new feature spec?"
"I need @support to help @sales with the customer issue"
"@legal @finance review this contract together"
```