create-tsi - Generative AI Toolkit for Creating AI Applications with Ease and Speed

Introduction to Create TSI

create-tsi is an innovative toolkit designed to streamline the creation of AI Applications using the LlamaIndex platform. With its focus on low-code solutions, create-tsi empowers users to effortlessly build sophisticated AI applications that leverage Large Language Models (LLMs) hosted by T-Systems on the Open Telekom Cloud.

Purpose and Functionality

The primary goal of create-tsi is to simplify, accelerate, and provide flexibility in the AI application development process. It allows users to generate bots, write custom agents, and tailor these applications for diverse use cases. This is particularly beneficial for individuals or organizations seeking to harness AI without engaging in extensive manual coding efforts.

Getting Started

Interested users will need a T-Systems API key to access create-tsi. A trial access request can be made through an online form. Once the API key is secured, setting up the application is as simple as running a command:

npx create-tsi@latest

This command initializes the application. Users can then follow the guidance in the generated README.md file to launch and operate their new AI app.

Features of the Generated Application

Next.js Frontend: The AI application comes equipped with a Next.js-powered frontend, designed to function as a chat interface. This interface can competently respond to user inquiries regarding the data provided to it.
Python FastAPI Backend: The backend is developed using Python’s FastAPI framework and is integrated with the llama-index python package. This environment ensures robust backend support for AI operations and interactions.
Data Handling: The application can index and respond to questions about user-supplied data. Users can populate the data directory within the backend folder with supported file types. The app then processes and utilizes this data in generating responses.

Example Usage

To utilize create-tsi, users can start by running it in interactive mode, selecting options, and providing necessary configuration details. During this setup, users will choose the project name, decide on a Next.js frontend, enter their T-Systems API key, select a model, and specify a data source. This interactive setup is flexible, allowing for various paths according to individual requirements or preferences.

Code of Conduct and Licensing

create-tsi maintains a strict adherence to community standards as outlined by the Contributor Covenant Code of Conduct. All participants in the project are expected to abide by these guidelines.

The project also adheres to the REUSE standard for software licensing, ensuring clarity and accessibility in licensing information. Each file in the project includes its respective copyright and license details, adhering to the MIT License format.

Conclusion

create-tsi is a user-friendly, low-code AI tool that simplifies the creation and customization of AI applications. By providing powerful backend support and an intuitive frontend interface, it serves as a valuable asset for those looking to integrate AI solutions into their operations swiftly and efficiently.