How to install SearchAI ChatBot

SearchAI ChatBot — Starting a whole new conversation

A conversational AI chatbot trained on your custom data sources. That’s what SearchAI ChatBot is. Here’s how you can get it up and running — fast.

What is SearchAI ChatBot, and how does it help?

In today’s fast-paced digital world, your enterprise content, spread across your website, documents, and FAQs, holds a treasure trove of information. But it’s only as valuable as it is accessible.

SearchAI ChatBot unlocks value — transforming your static content into a dynamic, conversational AI assistant.

In addition, SearchAI has established industry-leading privacy and security standards. Unlike many other chatbot solutions, it runs entirely on your infrastructure, ensuring your data remains under your control.

How Does SearchAI ChatBot Work?

Imagine a chatbot that doesn’t just parrot pre-written responses but deeply understands your content and can answer questions in a natural, human-like way. That’s the power of SearchAI Chatbot.

It’s built on cutting-edge large language models (LLMs) like ChatGPT but trained specifically on your data sources. This means it can tap into your FAQs, manuals, and other documents to provide instant, contextually relevant answers. Plus, you retain complete control — you can review, refine, and customize the chatbot’s responses to ensure they’re always on point.

Ready to see it in action? Let’s dive into the installation process!

Installing SearchAI ChatBot — Step-by-Step

Understanding the SearchAI ChatBot System Architecture

Before you dive into deploying SearchAI ChatBot. Let’s look at three key components that work together to power the ChatBot:

SearchBlox Enterprise Search

The user interface where you create, manage, and interact with your chatbots.

SearchAI Chatbot Server

Handles communication between SearchBlox and the private LLM server and stores chatbot embeddings.

SearchAI Private LLM Server

The LLM is the brain responsible for generating responses to user queries using generative AI.

Prerequisites (System requirements and conditions before you begin)

Before you begin, ensure your infrastructure meets the following requirements:

Operating System: A Linux server running Red Hat/CentOS or Ubuntu.

Hardware Resources: Sufficient CPU, memory, and disk space to run the LLM model effectively. Refer to the SearchBlox documentation for detailed recommendations based on your anticipated usage.

Software: Docker Desktop installed and running. Download it from the Docker website.

Files: You’ll need the following files:

1. image.tar (Chatbot server image)
2. docker-compose-prod.yml (configuration file for Docker)
3. ChatBotServerFiles.zip (additional server files)
4. privatellm-package-v0.9.zip (for private LLM server testing)

Note : SearchBlox support will provide these files. You must create a support ticket , and the SearchBlox support team will get back to you.

Step 1: Deploy SearchBlox Enterprise Search

Let’s begin by deploying your SearchBlox Enterprise Search instance. As mentioned earlier, this would be the interface where you can create and manage your SearchAI ChatBot.

Visit our detailed support page, where we document how to install Enterprise Search across operating systems.

Step 2: Chatbot Server Deployment

1. Load Docker Images

Use the following commands to load the chatbot’s required components into Docker.

docker image load --input chatbot2-api.tar
docker image load --input redis.tar
docker image load --input chatbot_docker-alb.tar

Remember: These credentials are critical for secure communication with your SearchBlox server. It’s essential to keep them confidential and avoid storing them in plain text wherever possible.

2. Update Configuration (.env) File

· Go to the chatbot2_docker folder on your server.

· Open the .env file with a text editor.

· Locate the section titled “#SearchBlox Server Settings”. Here, you’ll need to update the placeholders (SBX_URL, SBX_USERNAME, and SBX_PASSWORD) with your OpenSearch credentials.

Note: If you have trouble finding the required OpenSearch credentials (SBX_URL, SBX_USERNAME, and SBX_PASSWORD), create a support ticket to reach out to the SearchBlox Support team.

3. Start the Chatbot Server

In your terminal, navigate to the chatbot2_docker directory.

cd chatbot2_docker

Run this command to start the server

nohup docker compose -f docker-compose-prod.yml up &

Step 3: Install the Private LLM (the Chatbot’s “Brain”)

Red Hat/CentOS Instructions

1. Install OpenJDK 17

Use the following commands to load the chatbot’s required components into Docker.

sudo yum install java-17-openjdk -y

2. Install System Monitoring Tool (Optional)

sudo dnf copr enable atim/bottom -y
sudo dnf install bottom -y

3. Install “wget” and “unzip”

sudo dnf copr enable atim/bottom -y
sudo dnf install bottom -y

Set Up the LLM

1. Navigate to the installation directory

cd /opt/

2. Unzip the LLM package (privatellm19012024.zip) provided by SearchBlox support into the /opt/ directory

unzip privatellm19012024.zip

3. Create a user named “searchai”

useradd searchai

4. Create a service file named privatellm.service in /etc/systemd/system/

Paste the provided service file content into the file.

sudo vi /etc/systemd/system/privatellm.service

Ubuntu Instructions

If you have already installed the LLM server on Red Hat/CentOS, you can skip this step and move on to the “Start the Service” step.

1. Install OpenJDK 17

Use the following commands to load the chatbot’s required components into Docker.

sudo apt install openjdk-17-jdk -y

2. Install System Monitoring Tool (Optional)

sudo apt update -y
sudo apt install bottom -y

3. Install “unzip”

sudo apt-get install unzip -y

Set Up the LLM

1. Follow the steps in “Setting Up the LLM” for RedHat/CentOS.
2. Instead of creating the service file under /etc/systemd/system/, create it under /etc/system/system/:Navigate to the installation directory:

sudo vi /etc/system/system/privatellm.service

Paste the provided service file content into the file.

Step 4. Start the Services (For Both Red Hat/CentOS and Ubuntu)

1. Start the CPP Server

systemctl start cppserver

2. Start the Private LLM Server

systemctl start privatellm

Note: Both the cppserver and privatellm services must be running simultaneously for the chatbot to function.

Now, your SearchAI ChatBot should be functional. Reach out to our support team if you need help following these steps.

Next Steps — Getting started with SearchAI ChatBot

For a comprehensive guide on using the app, including creating and managing chatbots, refer to our detailed support page.

SearchBlox support team will assist you with any questions or issues you may have encountered during the installation process.