LANGCHAIN — Can AI Companions Be Created by RealChar X LangSmith?

The most profound technologies are those that disappear. They weave themselves into the fabric of everyday life until they are indistinguishable from it. — Mark Weiser.

Creating AI companions using RealChar and LangSmith is an exciting endeavor. By leveraging open source tools in the Generative AI/LLM space, including LangChain, developers can build sophisticated AI companions. Below, we will explore how RealChar utilized LangSmith and LangChain to monitor, observe, and evaluate the traffic and conversations with their AI companion.

Using LangSmith for Monitoring and Observability

RealChar found LangSmith to be an essential tool for monitoring and observing the traffic and conversations with their AI companion. LangSmith’s logging and filtering capabilities allowed them to track issues accurately. The following code snippet demonstrates how to set up the required environment variables for using LangSmith:

LANGCHAIN_TRACING_V2=true
LANGCHAIN_ENDPOINT=https://api.smith.langchain.com
LANGCHAIN_API_KEY=YOUR_LANGCHAIN_API_KEY
LANGCHAIN_PROJECT=YOUR_LANGCHAIN_PROJECT

Evaluating Conversations and Adding to Dataset

In addition to monitoring, LangSmith also facilitated the identification of important conversations and adding them to the dataset for evaluation. RealChar found this feature helpful for evaluating and safe checking the prompts going forward. The following code snippet showcases how to identify important conversations and add them to the dataset using LangSmith:

// Identify important conversation
const importantConversation = ...

// Add to dataset
LangSmith.addtoDataset(importantConversation);

Utilizing LangSmith for Analytics

Moreover, RealChar highlighted the usefulness of LangSmith for analytics, observability, and evaluation, all in one place. They emphasized that LangSmith is very useful for a production-level application with a large volume of traffic like RealChar. The integration with LangChain made it almost free to access these features.

Conclusion

In summary, RealChar’s experience with LangSmith and LangChain showcases the power of open source tools in creating AI companions. By leveraging LangSmith for monitoring, observability, and evaluation, developers can build and maintain sophisticated AI companions with ease.

In this tutorial, we explored how RealChar utilized LangSmith for monitoring, observing, and evaluating conversations with their AI companion. The provided code snippets demonstrate the ease of integrating LangSmith with existing projects. With the power of LangChain and LangSmith, developers can create highly functional and efficient AI companions.