A wrapper around the Perplexity WebUI using Langchain. Under the hood, relies on reteps/perplexityai.
from perplexity_webui_langchain import PerplexityWebUIChatModel
from langchain_core.messages import HumanMessage
import asyncio
llm = PerplexityWebUIChatModel(email="peteras4@illinois.edu")
async def main():
msg = await llm.ainvoke([
HumanMessage(content="What is the capital of Illinois?"),
])
print(msg.content)
async for msg in llm.astream("hello"):
print(msg.content, end="|", flush=True)
async for msg in llm.astream("what was my last message?", followup=True):
print(msg.content, end="|", flush=True)
if __name__ == "__main__":
asyncio.run(main())python3 -m pip install git+https://github.com/reteps/perplexity-webui-langchain.git- The perplexity pro plan supports custom LLMs (e.g. claude3.5) on the WebUI (and can be scripted through a WebSocket)
- I want to be able to use these custom LLMs in other settings, like autocomplete in my code editor
- I can hook up Langchain as an OpenAI chat completions compatible endpoint using this repo
