LangSmith Fetch: A New CLI Tool for Seamless Agent Debugging

LangChain has announced the release of LangSmith Fetch, an innovative command-line interface (CLI) tool designed to streamline the debugging process for developers working with agents. This tool integrates the comprehensive capabilities of LangSmith tracing directly into the terminal and integrated development environments (IDEs), enhancing the efficiency of agent debugging workflows.

Addressing Developer Frustrations

LangSmith Fetch aims to eliminate the common friction developers encounter when debugging agents. Traditionally, developers using coding tools such as Claude Code or Cursor needed to switch to the LangSmith UI to access trace data, disrupting their workflow. With LangSmith Fetch, developers can now access trace data directly from the terminal, reducing the need for context-switching and enhancing productivity.

Features and Functionalities

The tool is built around two primary developer workflows: immediate trace fetching and bulk data export. Developers can retrieve the most recent trace data with a simple command, bypassing the need for web interfaces and manual data handling. For instance, the command langsmith-fetch traces --project-uuid --format json provides instant access to trace data in JSON format.

LangSmith Fetch also supports bulk export functionalities, allowing developers to extract large datasets for evaluation and analysis. This feature is particularly beneficial for creating test suites and analyzing production failures.

Empowering Coding Agents

One of the standout features of LangSmith Fetch is its ability to enhance the capabilities of coding agents. By providing direct access to comprehensive agent execution data, coding agents like Claude Code can analyze decision-making processes, identify inefficiencies, and suggest improvements based on real execution data. This integration transforms coding agents into powerful debugging tools.

Integration and Compatibility

LangSmith Fetch seamlessly integrates with existing LangSmith setups without requiring additional configuration. Developers can install the tool via pip and begin utilizing its features immediately. The tool uses the same authentication protocols as the LangSmith platform, ensuring a smooth transition for current users.

Why a CLI Tool?

The decision to develop a CLI tool, rather than an MCP server, was driven by the need for flexibility and compatibility with existing developer workflows. While MCP (Model Context Protocol) excels in real-time tool calling, a CLI tool offers greater versatility, allowing developers to integrate it into various workflows and tools without restrictions.

LangChain emphasizes the CLI’s alignment with the Unix philosophy of flexibility and composability, enabling developers to leverage LangSmith Fetch in numerous ways, from quick trace inspections to batch processing and integration with coding agents.

For more information and to get started with LangSmith Fetch, visit the LangChain Blog.

Image source: Shutterstock