Deploy

Overview

This guide provides step-by-step instructions for setting up an environment that integrates multiple services like AIMETA.

Prerequisites

• Necessary API keys and environment variables as required by specific services.

• Ensure NVIDIA GPU drivers are installed if utilizing GPU support for the LLM.

Service Configuration

1. LLM-GPU Services

   • Purpose: Hosts a Large Language Model with optional GPU support.

   • Configuration:

     • Operates within a Linux environment.

     • Utilizes all available NVIDIA GPUs for the GPU-enabled configuration.

2. Pull-Model Service

   • Purpose: Manages the model pulling for the LLM service.

   • Configuration:

     • Built from Linux environment.

     • Depends on environment variables such as model URLs and API keys.

     • Connected to the net network and dependent on llm-gpu.

3. Database Service

   • Purpose: Utilizes Neo4j for graph-based data storage.

   • Configuration:

     • Exposes ports for external connectivity.

     • Uses persistent storage volumes.

     • Configured with environment variables for authentication and plugins.

4. Loader Service

   • Purpose: Loads data into the Neo4j database.

   • Configuration:

     • Custom Dockerfile.

     • Interacts with Neo4j and depends on the database and pull-model services.

     • Utilizes API keys and environment configurations.

5. Bot Services

   • General Bot Configuration:

     • Custom Dockerfile.

     • Utilizes net network and Neo4j database connections.

   • Voice Bot Specifics:

     • Handles tasks related to Voice processing.

6. API Service

   • Purpose: Provides an API interface for external communications.

   • Configuration:

     • Built from a custom Dockerfile.

     • Health checks implemented.

     • Ports exposed for external connectivity.

     • Depends on the database and pull-model service completion.

Health Checks

• Implemented in services such as the API and chrome plugin to ensure operational status before being available for use.