---
type: "post"
id: "debug-long-running-agents-faster-skip-the-reconstruction-9f97f6"
slug: "debug-long-running-agents-faster-skip-the-reconstruction-9f97f6"
title: "LangChain has launched Trajectories in LangSmith, a chronological, conversational view of agent sessions"
description: "Thesis: Trajectories make the session’s behavioral path a usable unit for debugging and review, while leaving the underlying trace as the execution-detail record."
retrieval_nugget: "Thesis: Trajectories make the session’s behavioral path a usable unit for debugging and review, while leaving the underlying trace as the execution-detail record."
published_at: "2026-09-29"
updated_at: "2026-09-29"
record_date: "2026-09-29"
date_kind: "published_at"
topics: []
source_urls: ["https://www.langchain.com/blog/langsmith-trajectories-tracing"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:debug-long-running-agents-faster-skip-the-reconstruction-9f97f6"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:0897d3d04636130b2017e1961071050dfd273dec60cf7ad64cff9c931eb428f6","reviewed_at":"2026-09-29T20:07:13.070Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Trajectories make the session’s behavioral path a usable unit for debugging and review, while leaving the underlying trace as the execution-detail record.","observed_facts":[{"text":"LangChain describes Trajectories as aggregating messages from humans, AI, and tools across a main agent and its subagents, in the order they first appeared, with each message shown once.","source_urls":["https://www.langchain.com/blog/langsmith-trajectories-tracing"]},{"text":"The company says trajectories can be scored by online evaluators and routed to annotation queues or datasets.","source_urls":["https://www.langchain.com/blog/langsmith-trajectories-tracing"]},{"text":"LangChain says the feature is available on all LangSmith plans in the US.","source_urls":["https://www.langchain.com/blog/langsmith-trajectories-tracing"]}],"mechanism":"A trajectory is a projection over traces in a thread: it removes nested run structure while retaining the messages and actions that explain behavior. This makes a long session easier to scan for the point of deviation, after which an engineer can inspect the corresponding full trace for inputs, outputs, timing, retries, and nested execution details.","why_now":"LangChain is presenting session-level behavior as an input to production evaluation and agent improvement, not merely as a more readable debugging display.","implications":["Reviewers who are not runtime engineers may be able to assess behavior through a readable session view.","Teams may also be able to connect production sessions to evaluator scoring, human feedback, datasets, and supervised fine-tuning examples."],"evidence_boundary":"These capabilities and benefits are described by LangChain in its product announcement; the supplied source does not provide independent measurements of debugging speed, evaluator quality, or downstream model improvement.","watch_conditions":["Look for evidence of how teams use trajectories in production, including whether online evaluators and human review produce reliable signals and whether exported sessions improve later testing or fine-tuning."],"related_records":[],"new_branch_reason":"No prior New Runtime record is supplied; this decision begins a distinct analytical record for LangSmith Trajectories and its session-level view of agent behavior."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/debug-long-running-agents-faster-skip-the-reconstruction-9f97f6/","markdown":"https://newruntime.com/posts/debug-long-running-agents-faster-skip-the-reconstruction-9f97f6.md","json":"https://newruntime.com/posts/debug-long-running-agents-faster-skip-the-reconstruction-9f97f6.json"}
---

# LangChain has launched Trajectories in LangSmith, a chronological, conversational view of agent sessions

## Retrieval answer

Thesis: Trajectories make the session’s behavioral path a usable unit for debugging and review, while leaving the underlying trace as the execution-detail record.

Thesis: Trajectories make the session’s behavioral path a usable unit for debugging and review, while leaving the underlying trace as the execution-detail record.

Observed facts: LangChain describes Trajectories as aggregating messages from humans, AI, and tools across a main agent and its subagents, in the order they first appeared, with each message shown once. The company says trajectories can be scored by online evaluators and routed to annotation queues or datasets. It also says the feature is available on all LangSmith plans in the US.

Mechanism: A trajectory is a projection over traces in a thread: it removes nested run structure while retaining the messages and actions that explain behavior. This makes a long session easier to scan for the point of deviation, after which an engineer can inspect the corresponding full trace for inputs, outputs, timing, retries, and nested execution details.

Why now: LangChain is presenting session-level behavior as an input to production evaluation and agent improvement, not merely as a more readable debugging display.

Implications: Reviewers who are not runtime engineers may be able to assess behavior through a readable session view. Teams may also be able to connect production sessions to evaluator scoring, human feedback, datasets, and supervised fine-tuning examples.

Evidence boundary: These capabilities and benefits are described by LangChain in its product announcement; the supplied source does not provide independent measurements of debugging speed, evaluator quality, or downstream model improvement.

Watch conditions: Look for evidence of how teams use trajectories in production, including whether online evaluators and human review produce reliable signals and whether exported sessions improve later testing or fine-tuning.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)