Skip to main content

Getting started with PhariaAI applications

Before looking at the process of developing and deploying a PhariaAI application, it is essential to understand what a PhariaAI application is. For a comprehensive overview, refer to What is a PhariaAI application?

If you're looking for a more hands-on approach, start with our Quick Start Guide to get your PhariaAI application up and running quickly.

Developing and deploying a PhariaAI application follows an iterative process similar to other software development workflows. While the phases below are presented sequentially, you can adapt and revisit them as needed based on your application's requirements. This section provides an overview of the key development phases and points you to detailed documentation for each step.

1. Infrastructure prerequisites​

To get started with PhariaAI application, ensure that you have installed PhariaAI with the following core components enabled:

  • PhariaEngine - For building and deploying skills
  • PhariaAssistant - For rendering your application UI
  • PhariaOS Manager - For deployment and management of your application
  • Pharia Data Platform - For data management utilities such as file management, file transformation, and so on
  • PhariaInference API - For model serving and inference. See our model recommendations guide to help you choose the right model for your use case.
  • Document Index - For creating document collections

2. Data setup, processing and management​

When building a PhariaAI application you will need to setup and process some data whether its for model training, creation of collections for document search, or experimentation during development of your skills and Applications. We provide Pharia Data Platform to help you setup and process data for your application, including file ingestion for dataset creation, other file transformations, and more.

We also provide a Vector Database, called Document Index, to support Retrieval Augmented Generation (RAG) setups. With Document Index, you can create collections with metadata and leverage powerful semantic and hybrid search functionality to retrieve relevant documents and contexts for your AI applications.

You can take a look at our How to set up collections guide to learn more about how to use the Document Index API to create collections.

3. Skill development and deployment​

PhariaEngine is designed to facilitate the development and deployment of your skills. For detailed information on the API specifications, refer to the PhariaEngine documentation here. To explore the integration of a skill into a PhariaAI application, follow the steps outlined here to create a base PhariaAI application that includes a skill.

4. Development and deployment of the application service and UI​

To initiate the development and deployment of your first PhariaAI application, go here for a streamlined process. This guide provides a comprehensive walkthrough to get you started efficiently.

5. Tracing, debugging and evaluating the application and skills​

You can do tracing, debugging, and evaluation of your skills and application service during development and also in production by using PhariaStudio. We provide a user interface and an SDK to help you do this. To get started, refer to our PhariaStudio Introduction.

6. Application feedback​

Please consult this page on how to store application feedback in Assistant API's database.

Troubleshooting guide​

For any issues you encounter during the development and deployment process, refer to our Troubleshooting Guide for common error solutions and best practices.

Next Steps​

PhariaAI is still evolving. Your feedback is invaluable. If you have any suggestions or comments, contact using the Aleph Alpha Support Portal.