Deployed at: https://www.textgen.dev/

TextGen

A web app and backend server that uses NLP ML algorithms to auto-generate a paragraph of text from a short supplied sentence start.

How it works

1. Web app - written in React, runs on the client. This has an input text box in which a user can place a sentence start. On pressing 'Generate' it sends the text to the backend Flask Python server
2. Flask server - has a /predict endpoint that accepts POST requests. Uses the NLP model to run inference on the text and generate the remainder of the paragraph
3. Model - based on simple high level Huggingface pipeline models

Docker

The app, model and server have been dockerized for use of deployment. AWS ECS + ECR used for deploying.