What is Callback in Langchain: Explanation and Example

In Langchain, a callback is a function or object that runs automatically during or after a process, like when a language model generates text. It helps you track progress, log events, or modify behavior without changing the main code flow.

How It Works

Think of a callback in Langchain like a helper that listens quietly while your main program runs. When something important happens, like the language model finishing a task, the callback wakes up and does its job, such as logging the result or updating a progress bar.

This is similar to how a friend might remind you when a timer goes off while you are cooking. You don’t have to stop what you’re doing; the callback just reacts when needed. In Langchain, callbacks let you add extra features like tracking or debugging without changing the core logic.

Example

This example shows how to use a simple callback in Langchain to print messages when the language model starts and finishes generating text.

python

from langchain.callbacks.base import BaseCallbackHandler
from langchain.llms import OpenAI

class PrintCallback(BaseCallbackHandler):
    def on_llm_start(self, serialized, prompts, **kwargs):
        print("LLM started generating text...")

    def on_llm_end(self, response, **kwargs):
        print("LLM finished generating text.")

llm = OpenAI(callbacks=[PrintCallback()])
result = llm("Say hello in a friendly way.")
print("Output:", result)

Output

LLM started generating text...
LLM finished generating text.
Output: Hello! How can I help you today?

When to Use

Use callbacks in Langchain when you want to monitor or react to events during language model operations without interrupting the main process. For example, you can log inputs and outputs for debugging, update user interfaces with progress, or collect usage statistics.

Callbacks are helpful in real-world applications like chatbots, where you want to show typing indicators or save conversation history automatically. They keep your code clean by separating core logic from side tasks.

Key Points

• Callbacks run automatically at specific events in Langchain processes.
• They help track progress, log data, or modify behavior without changing main code.
• Callbacks keep your code organized by separating side tasks from core logic.
• They are useful for debugging, UI updates, and analytics in language model applications.

Key Takeaways

Callbacks in Langchain let you run extra code automatically during language model events.

They help track, log, or react without changing your main program flow.

Use callbacks to keep your code clean and add features like progress updates or debugging.

Callbacks are essential for building interactive and maintainable language model applications.