Practical overview
When it fits
Use Create Tsi when you want to generate a configurable chat application and choose its interface, models and data source during setup.
Who it is for
- Developers who want a generated starting point for a chat application grounded in their data.
- Teams evaluating a configurable LlamaIndex-based application with a Next.js frontend and FastAPI backend.
Documented uses
- Generate a chat application that answers questions about selected documents or folders.
- Create a website-content question-answering app, with Chrome available for ingestion.
- Scaffold a chat application connected to a MySQL data source.
- Generate a simple chat without selecting a data source.
Requirements & limits
Generation requires a T-Systems API key. The README says trial access can be requested through a form.
- Generated apps can combine a Next.js frontend with a Python FastAPI backend; the frontend is a setup choice.
- The generator offers model and embedding-model choices during interactive setup.
- Available data-source choices include no data, an example PDF, supported local files, local folders, website content and MySQL.
- Website-content ingestion requires Chrome.
- The generated backend uses LlamaIndex, while the LLMs are hosted by T-Systems on Open Telekom Cloud.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Generate a data-grounded chat app
Scaffold a chat application configured to answer questions using a selected data source.
Before you start
- A T-Systems API key; the project says trial access can be requested through its form.
- Node.js package tooling to invoke the generator.
- For website-content ingestion, Chrome is required.
Steps
Run the command below in interactive mode.
npx create-tsi@latestEnter a project name and choose whether to generate the Next.js frontend.
Provide the T-Systems API key, then select a model, embedding model and data source.
Read the generated README.md for instructions to start the application.
Expected outcome
A generated application can include a Next.js chat interface and Python FastAPI backend, configured for the chosen data source.
Guide limitations
- The README does not give the generated application's start commands in the cited setup instructions.
- Only supported files are ingested; the README lists PDF, DOC, DOCX, XLS, XLSX and CSV as examples.
- Website-content ingestion requires Chrome.
Set up a local-files question-answering app
Generate an application configured to index supported files or folders provided locally.
Before you start
- A T-Systems API key.
- Node.js package tooling.
- Supported local files or a folder of data to provide to the generated backend.
Steps
Run the command below.
npx create-tsi@latestChoose local files or local folders when prompted for a data source.
After generation, place supported files in the generated
backend/datafolder.Follow the generated README.md to start the app.
Expected outcome
The generated application is configured to ingest supported local data and answer questions about it through chat.
Guide limitations
- The generated README provides the next-stage startup instructions.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 17 October 2026
- Original website title
- telekom/create-tsi · Official homepage title · retained in its source language