GoQuant OTC Trading Bot

A voice-based OTC (Over-The-Counter) trading assistant built with Next.js frontend and Go backend. The bot allows users to place simulated trading orders through voice commands using speech recognition and text-to-speech capabilities.

Architecture

• Frontend: Next.js 15 with TypeScript, Tailwind CSS
• Backend: Go with Gin framework
• Speech Recognition: Web Speech API
• Text-to-Speech: Web Speech Synthesis API
• Price Data: CoinGecko API

Features

• Voice-based trading interface
• Real-time speech recognition
• Text-to-speech responses
• Support for multiple exchanges (OKX, Bybit, Deribit, Binance)
• Real-time cryptocurrency price fetching
• Session management
• Debug console for troubleshooting

Prerequisites

• Node.js (v18 or higher)
• Go (v1.24.1 or higher)
• Modern web browser with Web Speech API support (Chrome, Edge, Safari)
• Microphone access for voice features

Installation

1. Clone the Repository

2. Backend Setup

Navigate to the backend directory:

cd backend

Install Go dependencies:

go mod download

Create a .env file in the backend directory:

touch .env

Add the following environment variables to .env:

BLAND_API_KEY=your_bland_api_key_here
PORT=8080

3. Frontend Setup

Navigate to the frontend directory:

cd ../quantbot

Install dependencies:

npm install

Create a .env.local file in the quantbot directory:

touch .env.local

Add the following environment variable:

BACKEND_URL=http://localhost:8080

Running the Application

1. Start the Backend Server

In the backend directory:

go run main.go

The backend will start on http://localhost:8080

You should see:

Server starting on port 8080

2. Start the Frontend Development Server

In a new terminal, navigate to the quantbot directory:

cd quantbot
npm run dev

The frontend will start on http://localhost:3000

3. Access the Application

Open your browser and go to http://localhost:3000

🎤 Using the Trading Bot

Initial Setup

1. Allow microphone access when prompted by your browser
2. Test your microphone using the "Test Mic" button
3. Start a call by clicking the "Start Call" button

Voice Commands Flow

1. Choose Exchange: Say "OKX", "Bybit", "Deribit", or "Binance"
2. Select Symbol: Say the trading symbol (e.g., "Bitcoin", "BTC", "Ethereum")
3. Specify Order Details: Say quantity and price (e.g., "1.5 Bitcoin at 65000 dollars")
4. Confirm Order: Say "yes" to confirm or "no" to modify

Example Conversation

Bot: "Hello! Welcome to GoQuant's OTC trading service. Please choose an exchange: OKX, Bybit, Deribit, or Binance."
You: "OKX"
Bot: "Great! You've selected OKX. Which trading symbol would you like to trade?"
You: "Bitcoin"
Bot: "The current price for BITCOIN on OKX is $65,123.45. Now, what quantity and price for the order?"
You: "1.5 Bitcoin at 65000 dollars"
Bot: "Got it. To confirm, you want to trade 1.5000 BITCOIN at $65000.0000 per unit on OKX. Is that correct?"
You: "Yes"
Bot: "Excellent! Your simulated order has been recorded. Thank you for using GoQuant!"

Troubleshooting

Common Issues

1. Microphone Not Working

• Ensure microphone permissions are granted
• Check browser compatibility (Chrome, Edge, Safari recommended)
• Test microphone using the "Test Mic" button

2. Speech Recognition Errors

• Speak clearly and at a moderate pace
• Ensure good microphone quality
• Check for background noise

3. Backend Connection Issues

• Verify backend is running on port 8080
• Check that CORS is properly configured
• Ensure no firewall blocking the connection

4. Price Fetching Issues

• Check internet connection
• CoinGecko API might be rate-limited
• Fallback mock prices will be used if API fails

Debug Information

The application includes a debug console that shows:

• Speech recognition status
• Microphone permissions
• API calls and responses
• Error messages

API Endpoints

Backend Endpoints

• POST /start-call - Initialize a new trading session
• POST /bland/webhook - Process voice commands
• GET /health - Health check endpoint

Frontend API Routes

• POST /api/start-call - Proxy to backend start-call
• POST /api/process-speech - Proxy to backend webhook

Project Structure

├── backend/
│   ├── handlers/
│   │   └── bland.go          # Main trading logic
│   ├── main.go               # Server entry point
│   ├── go.mod                # Go dependencies
│   └── .env                  # Environment variables
├── quantbot/
│   ├── app/
│   │   ├── api/              # Next.js API routes
│   │   ├── globals.css       # Global styles
│   │   ├── layout.tsx        # Root layout
│   │   └── page.tsx          # Main application
│   ├── next.config.ts        # Next.js configuration
│   └── .env.local            # Environment variables
└── README.md

Security Notes

• This is a simulation/demo - no real trades are executed
• Microphone access is required for voice features
• All trading data is stored in memory (sessions are not persistent)
• No sensitive trading credentials are handled

Deployment

Backend Deployment

• Set environment variables in your deployment platform
• Ensure PORT environment variable is set correctly
• Configure CORS for your frontend domain

Frontend Deployment

• Update BACKEND_URL in environment variables
• Ensure your deployment platform supports Web Speech API
• Configure proper HTTPS (required for microphone access in production)