Skip to main content

Log10

This page covers how to use the Log10 within LangChain.

What is Log10?​

Log10 is an open-source proxiless LLM data management and application development platform that lets you log, debug and tag your Langchain calls.

Quick start​

  1. Create your free account at log10.io
  2. Add your LOG10_TOKEN and LOG10_ORG_ID from the Settings and Organization tabs respectively as environment variables.
  3. Also add LOG10_URL=https://log10.io and your usual LLM API key: for e.g. OPENAI_API_KEY or ANTHROPIC_API_KEY to your environment

How to enable Log10 data management for Langchain​

Integration with log10 is a simple one-line log10_callback integration as shown below:

from langchain.chat_models import ChatOpenAI
from langchain.schema import HumanMessage

from log10.langchain import Log10Callback
from log10.llm import Log10Config

log10_callback = Log10Callback(log10_config=Log10Config())

messages = [
    HumanMessage(content="You are a ping pong machine"),
    HumanMessage(content="Ping?"),
]

llm = ChatOpenAI(model_name="gpt-3.5-turbo", callbacks=[log10_callback])

Log10 + Langchain + Logs docs

More details + screenshots including instructions for self-hosting logs

How to use tags with Log10​

from langchain.llms import OpenAI
from langchain.chat_models import ChatAnthropic
from langchain.chat_models import ChatOpenAI
from langchain.schema import HumanMessage

from log10.langchain import Log10Callback
from log10.llm import Log10Config

log10_callback = Log10Callback(log10_config=Log10Config())

messages = [
    HumanMessage(content="You are a ping pong machine"),
    HumanMessage(content="Ping?"),
]

llm = ChatOpenAI(model_name="gpt-3.5-turbo", callbacks=[log10_callback], temperature=0.5, tags=["test"])
completion = llm.predict_messages(messages, tags=["foobar"])
print(completion)

llm = ChatAnthropic(model="claude-2", callbacks=[log10_callback], temperature=0.7, tags=["baz"])
llm.predict_messages(messages)
print(completion)

llm = OpenAI(model_name="text-davinci-003", callbacks=[log10_callback], temperature=0.5)
completion = llm.predict("You are a ping pong machine.\nPing?\n")
print(completion)

You can also intermix direct OpenAI calls and Langchain LLM calls:

import os
from log10.load import log10, log10_session
import openai
from langchain.llms import OpenAI

log10(openai)

with log10_session(tags=["foo", "bar"]):
    # Log a direct OpenAI call
    response = openai.Completion.create(
        model="text-ada-001",
        prompt="Where is the Eiffel Tower?",
        temperature=0,
        max_tokens=1024,
        top_p=1,
        frequency_penalty=0,
        presence_penalty=0,
    )
    print(response)

    # Log a call via Langchain
    llm = OpenAI(model_name="text-ada-001", temperature=0.5)
    response = llm.predict("You are a ping pong machine.\nPing?\n")
    print(response)

How to debug Langchain calls​

Example of debugging

More Langchain examples