Trajectories now in LangSmith: A readable view of every agent session

LangChain has launched Trajectories in LangSmith, described as a chronological, conversational view of an agent session. A trajectory aggregates messages from humans, AI, and tools across the main agent and any subagents, then shows them in the order they first appeared.

In LangSmith, every unit of work an agent performs is recorded as a run. Runs for a single operation form a trace, and traces from a multi-turn session are linked into a thread. A trajectory is a projection over the traces in a thread: it removes the nested run structure and keeps the messages and actions that explain the agent's behavior, with each message appearing once, in order.

LangChain said trajectories work out of the box for traces sent from LangChain, LangGraph, and Deep Agents, from agent SDKs like OpenAI and Claude, and from coding agents like Codex, Claude Code, and Cursor.

Users can now view threads as trajectories, score trajectories with online evaluators, and route them to annotation queues or datasets. LangChain said online evaluators scoring trajectories give better input for judging agent behavior across a session, because a trajectory keeps each message once in order rather than repeating context turn over turn.

According to LangChain, trajectories include system prompts, user messages, assistant responses, tool calls, and tool outputs, which the company said makes them useful for reviewing final answers and for post-training work. High-quality trajectories can be saved to datasets and exported for supervised fine-tuning workflows.

Trajectories are available now on all plans in the US, according to LangChain.