WebSocket

Overview

WebSocket support enables real-time, bidirectional communication with AI providers. Use this for interactive applications that require low-latency streaming.

WebSocket support is currently available for relay functionality. Standard chat completions use Server-Sent Events (SSE) for streaming.

Endpoint

ws://localhost:18080/v1/ws

Upgrade from HTTP to WebSocket using standard WebSocket handshake.

Authentication

Query Parameter (Recommended)

const ws = new WebSocket(
  'ws://localhost:18080/v1/ws?apiKey=sk-test-123'
);

Sec-WebSocket-Protocol Header

const ws = new WebSocket(
  'ws://localhost:18080/v1/ws',
  ['access_token', 'sk-test-123']
);

Disable Authentication (Development Only)

config.yaml

websocket_auth: false  # Not recommended for production

Disabling WebSocket authentication allows unauthenticated access. Only use in development environments.

Message Protocol

Client to Server

Send JSON messages to the server:

{
  "type": "chat.message",
  "model": "gemini-2.5-pro",
  "messages": [
    {"role": "user", "content": "Hello!"}
  ],
  "stream": true
}

Server to Client

Receive streaming chunks or complete responses:

{
  "type": "chat.chunk",
  "content": "Hello",
  "done": false
}

{
  "type": "chat.chunk",
  "content": "!",
  "done": true
}

Examples

JavaScript/Browser

const ws = new WebSocket(
  'ws://localhost:18080/v1/ws?apiKey=sk-test-123'
);

ws.onopen = () => {
  console.log('Connected');
  
  // Send chat request
  ws.send(JSON.stringify({
    type: 'chat.message',
    model: 'gemini-2.5-pro',
    messages: [
      { role: 'user', content: 'Tell me a joke' }
    ],
    stream: true
  }));
};

ws.onmessage = (event) => {
  const data = JSON.parse(event.data);
  
  if (data.type === 'chat.chunk') {
    process.stdout.write(data.content);
    
    if (data.done) {
      console.log('\nComplete!');
      ws.close();
    }
  }
};

ws.onerror = (error) => {
  console.error('WebSocket error:', error);
};

ws.onclose = () => {
  console.log('Disconnected');
};

Python

import asyncio
import websockets
import json

async def chat():
    uri = "ws://localhost:18080/v1/ws?apiKey=sk-test-123"
    
    async with websockets.connect(uri) as ws:
        # Send request
        await ws.send(json.dumps({
            "type": "chat.message",
            "model": "gemini-2.5-pro",
            "messages": [
                {"role": "user", "content": "Tell me a joke"}
            ],
            "stream": True
        }))
        
        # Receive chunks
        async for message in ws:
            data = json.loads(message)
            
            if data["type"] == "chat.chunk":
                print(data["content"], end="", flush=True)
                
                if data["done"]:
                    print("\nComplete!")
                    break

asyncio.run(chat())

Node.js

import WebSocket from 'ws';

const ws = new WebSocket(
  'ws://localhost:18080/v1/ws?apiKey=sk-test-123'
);

ws.on('open', () => {
  console.log('Connected');
  
  ws.send(JSON.stringify({
    type: 'chat.message',
    model: 'gemini-2.5-pro',
    messages: [
      { role: 'user', content: 'Tell me a joke' }
    ],
    stream: true
  }));
});

ws.on('message', (data) => {
  const message = JSON.parse(data);
  
  if (message.type === 'chat.chunk') {
    process.stdout.write(message.content);
    
    if (message.done) {
      console.log('\nComplete!');
      ws.close();
    }
  }
});

ws.on('error', (error) => {
  console.error('WebSocket error:', error);
});

ws.on('close', () => {
  console.log('Disconnected');
});

Message Types

Request Types

Type	Description	Required Fields
`chat.message`	Send chat completion request	`model`, `messages`
`ping`	Keep-alive ping	None
`cancel`	Cancel ongoing request	`request_id`

Response Types

Type	Description	Fields
`chat.chunk`	Streaming content chunk	`content`, `done`
`chat.complete`	Full response	`content`, `usage`
`error`	Error occurred	`message`, `code`
`pong`	Ping response	`timestamp`

Configuration

Server Configuration

config.yaml

# WebSocket settings
websocket_auth: true
websocket_timeout: 300  # Seconds
websocket_max_message_size: 1048576  # 1MB

Custom Route

import (
    "github.com/traylinx/switchAILocal/internal/api"
    "github.com/traylinx/switchAILocal/internal/relay"
)

server := api.NewServer(cfg, authManager, accessManager, configPath, luaEngine)

// Custom WebSocket relay
relay := relay.NewRelay(server.GetHandlers())
server.AttachWebsocketRoute("/custom/ws", relay)

Advanced Features

Connection Management

class WebSocketManager {
  constructor(url, apiKey) {
    this.url = `${url}?apiKey=${apiKey}`;
    this.ws = null;
    this.reconnectAttempts = 0;
    this.maxReconnects = 5;
  }

  connect() {
    this.ws = new WebSocket(this.url);

    this.ws.onopen = () => {
      console.log('Connected');
      this.reconnectAttempts = 0;
    };

    this.ws.onclose = () => {
      console.log('Disconnected');
      this.reconnect();
    };

    this.ws.onerror = (error) => {
      console.error('Error:', error);
    };
  }

  reconnect() {
    if (this.reconnectAttempts >= this.maxReconnects) {
      console.error('Max reconnection attempts reached');
      return;
    }

    this.reconnectAttempts++;
    const delay = Math.min(1000 * Math.pow(2, this.reconnectAttempts), 30000);

    console.log(`Reconnecting in ${delay}ms...`);
    setTimeout(() => this.connect(), delay);
  }

  send(message) {
    if (this.ws && this.ws.readyState === WebSocket.OPEN) {
      this.ws.send(JSON.stringify(message));
    }
  }

  close() {
    if (this.ws) {
      this.ws.close();
    }
  }
}

const manager = new WebSocketManager(
  'ws://localhost:18080/v1/ws',
  'sk-test-123'
);
manager.connect();

Keep-Alive

Send periodic pings to maintain connection:

const PING_INTERVAL = 30000; // 30 seconds

const ws = new WebSocket(
  'ws://localhost:18080/v1/ws?apiKey=sk-test-123'
);

let pingInterval;

ws.onopen = () => {
  // Start ping interval
  pingInterval = setInterval(() => {
    if (ws.readyState === WebSocket.OPEN) {
      ws.send(JSON.stringify({ type: 'ping' }));
    }
  }, PING_INTERVAL);
};

ws.onclose = () => {
  clearInterval(pingInterval);
};

Request Cancellation

let currentRequestId = null;

function sendRequest(content) {
  currentRequestId = Date.now().toString();
  
  ws.send(JSON.stringify({
    type: 'chat.message',
    request_id: currentRequestId,
    model: 'gemini-2.5-pro',
    messages: [{ role: 'user', content }],
    stream: true
  }));
}

function cancelRequest() {
  if (currentRequestId) {
    ws.send(JSON.stringify({
      type: 'cancel',
      request_id: currentRequestId
    }));
  }
}

// Usage
sendRequest('Tell me a long story');

// Cancel after 2 seconds
setTimeout(() => cancelRequest(), 2000);

Error Handling

ws.onmessage = (event) => {
  const data = JSON.parse(event.data);
  
  if (data.type === 'error') {
    console.error(`Error ${data.code}: ${data.message}`);
    
    switch (data.code) {
      case 'auth_error':
        console.log('Authentication failed');
        break;
      case 'rate_limit':
        console.log('Rate limit exceeded');
        break;
      case 'provider_error':
        console.log('Provider unavailable');
        break;
      default:
        console.log('Unknown error');
    }
  }
};

Performance Considerations

Connection Pooling

Reuse WebSocket connections instead of creating new ones for each request:

// Good: Single persistent connection
const ws = new WebSocket(url);
// Reuse for multiple requests

// Avoid: New connection per request
function sendMessage(content) {
  const ws = new WebSocket(url); // Don't do this
  ws.send(content);
  ws.close();
}

Message Batching

Buffer and send messages in batches when possible:

const messageQueue = [];
let sendTimer = null;

function queueMessage(message) {
  messageQueue.push(message);
  
  if (!sendTimer) {
    sendTimer = setTimeout(() => {
      ws.send(JSON.stringify(messageQueue));
      messageQueue.length = 0;
      sendTimer = null;
    }, 100); // Batch every 100ms
  }
}

Compression

Enable WebSocket compression for large messages:

// Browser (automatic with permessage-deflate)
const ws = new WebSocket(url);

// Node.js with ws library
import WebSocket from 'ws';
const ws = new WebSocket(url, {
  perMessageDeflate: true
});

Troubleshooting

Connection Refused

Cause: WebSocket endpoint not available Solution:

Verify server is running: http://localhost:18080
Check WebSocket is enabled in config
Try HTTP endpoint first: /v1/models

Authentication Failed

Cause: Invalid or missing API key Solution:

Verify API key in query parameter or header
Check key is configured in config.yaml
Confirm websocket_auth: true if using authentication

Connection Timeout

Cause: Idle connection exceeded timeout Solution:

Implement periodic ping messages
Increase websocket_timeout in config
Enable automatic reconnection

Next Steps

Chat Completions

Use standard SSE streaming for chat

Authentication

Configure WebSocket authentication

Overview

Endpoints

Provider Formats

Management API

Overview

Endpoint

Authentication

Query Parameter (Recommended)

Sec-WebSocket-Protocol Header

Disable Authentication (Development Only)

Message Protocol

Client to Server

Server to Client

Examples

JavaScript/Browser

Python

Node.js

Message Types

Request Types

Response Types

Configuration

Server Configuration

Custom Route

Advanced Features

Connection Management

Keep-Alive

Request Cancellation

Error Handling

Performance Considerations

Troubleshooting

Connection Refused

Authentication Failed

Connection Timeout

Next Steps

Chat Completions

Authentication

​Overview

​Endpoint

​Authentication

​Query Parameter (Recommended)

​Sec-WebSocket-Protocol Header

​Disable Authentication (Development Only)

​Message Protocol

​Client to Server

​Server to Client

​Examples

​JavaScript/Browser

​Python

​Node.js

​Message Types

​Request Types

​Response Types

​Configuration

​Server Configuration

​Custom Route

​Advanced Features

​Connection Management

​Keep-Alive

​Request Cancellation

​Error Handling

​Performance Considerations

​Troubleshooting

​Connection Refused

​Authentication Failed

​Connection Timeout

​Next Steps

Chat Completions

Authentication

Overview

Endpoint

Authentication

Query Parameter (Recommended)

Sec-WebSocket-Protocol Header

Disable Authentication (Development Only)

Message Protocol

Client to Server

Server to Client

Examples

JavaScript/Browser

Python

Node.js

Message Types

Request Types

Response Types

Configuration

Server Configuration

Custom Route

Advanced Features

Connection Management

Keep-Alive

Request Cancellation

Error Handling

Performance Considerations

Troubleshooting

Connection Refused

Authentication Failed

Connection Timeout

Next Steps