langchain python
https://python.langchain.com/en/latest/ is a great library supports both js and python.
I prefer the python implementation.
There is a basic fastapi server (streaming support) example here https://github.com/hwchase17/chat-langchain
Is there a way to integrate the project as is to the wasp GPT SaaS template and have it deployed to the same server just binding to a different port?
6 Replies
Hmm you could potentially do that, but I made another wasp template that does document/embeddings search, just as that langchain repo does (the only difference is it doesn't send it to chat, but that wouldn't be hard to implement)
https://github.com/wasp-lang/starters/tree/main/embeddings
and I also used langchain extensively with Wasp here: https://github.com/vincanger/twitter-brainstorming-agent
Thanks @Vinny (@Wasp) yes saw the embeddings example will take a look at the twitter repo. Langchain is great, much more familiar with the python version though. I'll probably serve it as a separate API endpoint with fly
no problem. don't be afraid to try the js library though. very similar and with wasp operations it makes it easier than building a separate api, I'd imagine.
If you give it a try and have any Q's, just ask here 🙂
Thanks @Vinny (@Wasp) !!
Wohooo @shving90, you just became a Waspeteer level 1!
Will look into the js documentation