Promptly: A customizable ChatGPT API interface for OpenAPI models

Release v0.0.1 (What's new?).

Welcome to Promptly, an open-source research project aimed at interacting with OpenAPI compatible language models. If you’re an engineer, data scientist, or just an enthusiast looking for fine-grained control over your conversational model, you’re in the right place.

This project isn’t production-ready. It’s more like a sandbox where you can mess around with prompt injections, fine-tuning, and much more. Do expect bugs; after all, that’s part of the fun!

Features:

  • Customizable API Calls: Go beyond the traditional ChatGPT API limitations.

  • Prompt Injection: Want to steer the conversation? Inject prompts to guide the model.

  • Fine-Tuning Support: Got a model that’s been fine-tuned? This project has you covered.

  • Multiple Models: This isn’t a one-trick pony. Use any OpenAPI-compatible model you like.

Support

This project is a research tool and comes with no warranties. It might crash, produce nonsensical outputs, or accidentally start a thermonuclear war. Use at your own risk.

For any questions, remarks, or bug reporting, feel free to open an issue on GitHub.


Full Table of Contents

The User Guide

This part of the documentation, which is mostly prose, begins with some background information about Promptly, then focuses on step-by-step instructions for getting the most out of Promptly.

The Community Guide

This part of the documentation, which is mostly prose, details the Promptly ecosystem and community.

The Contributor Guide

If you want to contribute to the project, this part of the documentation is for you.

Project Information

Promptly is an open-source project under the MIT License, its documentation lives at Read the Docs, with its code available at GitHub. It’s rigorously tested on Python 3.10+ to ensure reliable interactions with ML models.

Contributions to Promptly are most welcome!

The API Documentation / Guide

If you are looking for information on a specific function, class, or method, this part of the documentation is for you.