Set up Falcon 2 TTS Server

The Falcon 2 TTS Server is the core engine of Murf’s on-prem text-to-speech system. It processes text input and generates speech locally, delivering fast, high-quality output without relying on external APIs. Running Falcon 2 within your own infrastructure ensures low-latency performance, data privacy, and full operational control.

Trial Deployment

This guide provides step-by-step instructions for deploying the Murf TTS (Text-to-Speech) service on an AWS EC2 instance using Docker with GPU acceleration.

Prerequisites

Hardware Requirements

Software Requirements

Operating System

• Ubuntu 22.04 LTS (recommended)
• Amazon Linux 2023 (supported)

Required Software

• Docker Engine (20.10.0 or later)
• NVIDIA Container Toolkit (nvidia-docker2)
• NVIDIA GPU Drivers (version 525.60.13 or later)
• AWS CLI (configured with appropriate IAM permissions)

Pre-deployment Checklist

NVIDIA GPU Drivers Installation

Check if drivers are already installed:

$
nvidia-smi

If you see GPU information displayed, drivers are already installed and you can skip to step 3.

If drivers are NOT installed, install them:

$
# Update package manager
$
sudo apt-get update
$
$
# Install NVIDIA drivers
$
$
sudo apt-get install -y nvidia-driver-535
$
$
# Reboot to load the driver
$
$
sudo reboot

$
nvidia-smi

$
# Add NVIDIA driver repository
$
sudo add-apt-repository ppa:graphics-drivers/ppa
$
sudo apt-get update
$
$
# Install latest driver
$
$
sudo apt-get install -y nvidia-driver-535
$
$
# Reboot
$
$
sudo reboot

Run nvidia-smi to verify the installation.

AWS GPU instances launched with Deep Learning AMIs or GPU-optimized AMIs usually have drivers pre-installed.

Docker Installation

Install Docker if not already installed:

$
# Update package manager (Ubuntu)
$
sudo apt-get update
$
$
# Amaxon linux
$
Sudo yum update
$
$
# Install Docker
$
sudo apt-get install -y docker.io
$
$
# Amazon linux
$
sudo dnf install docker -y
$
$
# Start and enable Docker
$
sudo systemctl start docker
$
sudo systemctl enable docker
$
$
# Add current user to docker group (optional, to run without sudo)
$
sudo usermod -aG docker $USER
$
$
newgrp docker

Nvidia Container Toolkit Installation

$
# Add NVIDIA package repositories
$
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
$
curl -s -L https://nvidia.github.io/libnvidia-container/gpgkey | sudo apt-key add -
$
curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \
>
  sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
$
$
# Install nvidia-docker2
$
sudo apt-get update
$
sudo apt-get install -y nvidia-docker2
$
$
# Restart Docker daemon
$
sudo systemctl restart docker

Verify GPU Access

Test that Docker can access the GPU:

$
sudo docker run --rm --gpus all nvidia/cuda:12.4.0-base-ubuntu22.04 nvidia-smi

You should see your GPU listed with driver information.

Get the Docker Image

Murf team will provide you with the Docker image for the Falcon 2 TTS Server. You will be provided with a docker pull command to pull the image.

Deployment Steps

$
sudo docker pull <docker_image_url>

$
# Required: Master secret for TTS authentication
$
export TTS_MASTER_SECRET="your-secure-secret-key-here"
$
$
# Optional: License Logic Agent (LLA) endpoint
$
# Default: http://localhost:8000
$
export LLA_ENDPOINT="http://localhost:8000"

$
sudo docker run -d \
>
    --gpus all \
>
    -e TTS_MASTER_SECRET="${TTS_MASTER_SECRET}" \
>
    -e LLA_ENDPOINT="${LLA_ENDPOINT}" \
>
    --name murf-tts \
>
    -p 80:8000 \
>
    --restart unless-stopped \
>
    <docker_image_url>

$
sudo docker ps

CONTAINER ID   IMAGE
1234567890ab   <docker_image_url>

Verification

Check the Container Logs

Monitor the startup logs to ensure the service initializes correctly:

$
sudo docker logs -f murf-tts

This should show the startup logs and the service should be ready to use.

Test the TTS Service

Option 1: Health Check via Browser Open your browser and navigate to:

http://<your-ec2-public-ip>/

You should see:

{"message": "Hello from Murf TTS!"}

Option 2: Interactive Audio Test Page Navigate to the test page:

http://<your-ec2-public-ip>/test-audio

This provides a web interface to:

• Paste JSON payloads
• Generate speech
• Play audio directly in the browser
• Download WAV files

Option 3: API Documentation

View the interactive API documentation:

http://<your-ec2-public-ip>/docs

This opens the FastAPI Swagger UI for exploring all available endpoints.