Skip to content

EVA WebSocket Chat API

Welcome to the EVA WebSocket Chat API documentation. EVA is a real-time chat API service built with WebSocket technology, featuring conversation history, intelligent summarization, and Retrieval-Augmented Generation (RAG) capabilities.

✨ Key Features

Real-time Communication : WebSocket-based instant messaging with low latency

Intelligent Responses : AI-powered responses with emotional context and segmented delivery

Conversation History : Persistent storage with automatic summarization for long conversations

RAG Integration : Retrieval-Augmented Generation with automatic document search and source attribution

Flexible Authentication : Multiple authentication methods to suit different platforms and use cases

Emotional Context : Responses include emotional indicators (neutral, happy, excited, thoughtful, etc.) for enhanced user experience

🚀 Quick Start

1. Get Your API Key

Contact your administrator to obtain your API key for authentication.

2. Choose Your Connection Method

const ws = new WebSocket('ws://your-domain.com/ws', ['chat', 'your-api-key']);

import websockets

extra_headers = {"Authorization": "Bearer your-api-key"}
ws = await websockets.connect('ws://api.example.com/ws', extra_headers=extra_headers)

wscat -c "ws://your-domain.com/ws?token=your-api-key"

3. Start Chatting

  1. Connect to the WebSocket endpoint
  2. Start a conversation by sending a start_conversation message
  3. Send messages and receive intelligent, emotionally-aware responses
  4. Enjoy the enhanced conversational experience!

📋 Message Flow

The API follows a simple but powerful message flow:

sequenceDiagram
    participant Client
    participant EVA API
    participant AI Engine
    participant Document Store

    Client->>EVA API: Connect with authentication
    EVA API-->>Client: Connection established

    Client->>EVA API: start_conversation
    EVA API-->>Client: conversation_started

    Client->>EVA API: user_message
    EVA API->>Document Store: Search relevant docs
    EVA API->>AI Engine: Generate response with context
    AI Engine-->>EVA API: Segmented response
    EVA API-->>Client: message (chunk 1, emotion: thoughtful)
    EVA API-->>Client: message (chunk 2, emotion: excited)
    EVA API-->>Client: message (chunk 3, is_final: true)

🎯 What Makes EVA Special

Emotional Intelligence

EVA doesn't just respond with text—it provides emotional context that helps create more engaging conversations. Whether the response is excited about a discovery or thoughtful about a complex topic, you'll know the tone.

Smart Segmentation

Long responses are intelligently broken into chunks, each with its own emotional context. This allows for progressive display and better user experience.

Automatic Context

The RAG system automatically searches and incorporates relevant documents into responses, with full source attribution so you know where information comes from.

Conversation Memory

Every conversation is preserved with intelligent summarization, so context is maintained across sessions without overwhelming the system.

🔗 Next Steps

Ready to build amazing conversational experiences with EVA? Let's get started!