A question about API #619
-
Thanks for creating such an amazing project! I noticed that in the main branch, the API is not implemented; in the dev branch, the API seems to be under development. If I want to use the API, should I switch to the dev branch to use it, or wait until the development in the dev branch is completed? |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
Thank you! Very happy you like it. I would say wait a couple days to a week, then use the API. It's been my goal to snapshot the gradio codebase, turn it into a release snapshot, then deprecate the Gradio version and focus solely on the API Server/clients. That said, I had thought it would be a bit faster turn around, but so it goes. I'm currently working to improve tests for the API to ensure stability/robustness, as I've finally gotten around to implementing a baseline RAG and embeddings creation system, and once I've cleaned it up and made it stable, will then, finally, merge the dev branch to main, fully deprecate Gradio, and encourage individuals to use the API/clients. The /media and /chat endpoints are currently stable with no planned changes until the next API version (i.e. not gonna change for a while). The RAG endpoint being an example, while its built out and somewhat works, I still need to do some thinking on how to structure the endpoint and compatibility with existing stuff, like other RAG solutions and openai's Files/RAG api, to make it easier for integration/compatibility with other software. |
Beta Was this translation helpful? Give feedback.
-
Thank you for your quick and detailed response! I'm trying to create an AI course video player based on your project, and I'm currently conducting further research into the structure of the project. I think the media endpoint currently meets my needs, and in the future, I might integrate RAG functionality – by then, it should be developed. |
Beta Was this translation helpful? Give feedback.
Thank you! Very happy you like it.
I would say wait a couple days to a week, then use the API. It's been my goal to snapshot the gradio codebase, turn it into a release snapshot, then deprecate the Gradio version and focus solely on the API Server/clients.
That said, I had thought it would be a bit faster turn around, but so it goes.
I'm currently working to improve tests for the API to ensure stability/robustness, as I've finally gotten around to implementing a baseline RAG and embeddings creation system, and once I've cleaned it up and made it stable, will then, finally, merge the dev branch to main, fully deprecate Gradio, and encourage individuals to use the API/clients.
The /media an…