Search code, repositories, users, issues, pull requests...

OpenGPTs is an open-source initiative with the goal of replicating the OpenAI "GPTs" product by creating open-source replicas of the several components that "GPTs" consist of. The successful execution of this project entails a fully observable and configurable GPTs alternative that can be deployed in a siloed environment, which can interface with any LLM, and can be extended to use unique tools.

Layers #3,4 will be highly modular so that they can serve a variety of other purposes while #1,2 are more interdependent and built specifically for the "GPTs" use case. This is a large project; you are welcome to contribute as much or as little as you want.

This project just recently was put into motion by a team of experienced developers, and we strive to make it fully open-source and will thus be following best practices to ensure an efficient development process. Hosting will be provided by Agent Artificial (an organization run by a subset of these developers).

This project is organized in a way so that there is a separation of responsibilities this enables systems to be independent and easier to develop. The OpenGPTs system as a whole begins with the Client & Business Layer API service where a user would create and use GPTs. All requests will be sent to the OS Assistants API which handles persistent storage and requesting generations, it functions in the same way as OpenAI's Assistants API. Finally, the request for generation is sent to Agent (Cognitive Architecture) which will handles the execution of the generation.

Leave a Comment