Skip to content
Live
Trending storyDeveloping

LangChain refactors LangSmith around Agent

1 reports1 outlets23 hours agoUpdate

The story so far

AI overview

LangChain announces it will refactor LangSmith around Agent, launching the LangSmith Agent Platform, unifying the four stages of building, running, improving, and governing into a single agent record. Previously, traces sent to LangSmith used a flat project hierarchy and were not associated with a specific agent. Traces of the same agent in different environments were also not associated. The new data model uses a single agent record, with traces grouped by environment. LangChain states that different types of agents require different evaluation methods, for example, coding agents check if the code is merged, while customer service agents require low latency. LangSmith uses a composable approach, allowing users to select features as needed. If an agent is already running on another platform, it can be integrated into LangSmith for improvement and governance without changing other workflows. The new version of LangSmith offers cloud, BYOC, and self-hosted environments globally; the general and coding types of agents are already available, while voice agents have not yet been launched; the centralized dashboard currently offers workspace-level access.

AI-written from the reports · Updated 17 minutes ago

Coverage timeline

Follow the coverage to see every side of the story.

17 Nov
  1. LangChain BlogFeatured
    We rebuilt LangSmith around the agent

    LangChain announced it will refactor LangSmith around Agents, launching the LangSmith Agent Platform, unifying the four stages of building, running, improving, and governing into a single agent record.

Heat over time

Not enough data yet to draw a trend.