I Get Different Responses in Langgraph Studio vs Running It Locally on VS Code

[Yogesh-Dubey-Ayesavi]

I have developed a configuration file that dynamically creates assistants and tools when passed to a service. While the same configuration works correctly in Visual Studio Code, I encounter errors when running it in Langgraph Studio.

My Graph

My Code

import asyncio
from colorama import Fore
from app.services.langchain_service import LangchainService
from app.utils.constants import sample
from app.utils.logger import setup_logging

setup_logging()


async def print_message(response):
    print(
        Fore.GREEN + f"Assistant: {response.text} Index -> {response.index}",
        flush=True,
    )


langchain_service = LangchainService(
    llm_config=sample.get("llm_config"),
    on_partial_response=print_message,
    assistants=sample.get("assistants"),
)

graph = langchain_service.graph

langchain_service.set_call_sid("1")

if __name__ == "__main__":
    try:

        async def main():
            while True:
                # Get input from the user in the console
                user_input = input(Fore.RED + "You: ")

                # Exit loop if user types 'exit'
                if user_input.lower() == "exit":
                    print("Exiting chat.")
                    break

                # Stream response from langchain service
                await langchain_service.run(user_input, stream_output=True)

        # Run the main event loop
        asyncio.run(main())
    except KeyboardInterrupt:
        print("Exiting chat ...")

When running in VS Code

When running in LangGraph Studio