Many users praised LangSmith's Codex tracing plugin and agent support for enabling production-grade debugging and observability across tools like Cursor and Copilot, while one dismissed it as inferior to a competitor.
Based on 5 visible X reactions from 8 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Useful direction. This gets much more valuable when teams can compare failures across Cursor, Copilot, and in-house agents with the same trace schema, then spot where tool calls or long sessions start to drift. The big win is not just visibility, but faster debugging and cheaper iteration inside the real developer workflow.
@LangChain This becomes valuable when traces stop being a demo artifact and become part of the daily debugging loop. Seeing tool calls, token burn, and subagent branches in one place should make it much easier to find where agent workflows actually break in production.
@LangChain This makes agent observability production-grade. I’d add three defaults: redact secrets before export, sample healthy runs, and retain 100% of failures. Then compare tool retries, recovery paths, latency, and cost per completed task across sessions.
@LangChain Yea its already built here and better http://www.martinloop.com
@LangChain this is huge for debugging, congrats team
@LangChain This becomes valuable when traces stop being a demo artifact and become part of the daily debugging loop. Seeing tool calls, token burn, and subagent branches in one place should make it much easier to find where agent workflows actually break in production.
@LangChain This makes agent observability production-grade. I’d add three defaults: redact secrets before export, sample healthy runs, and retain 100% of failures. Then compare tool retries, recovery paths, latency, and cost per completed task across sessions.
@LangChain this is huge for debugging, congrats team
we did a bunch of work to standardize how coding agents trace to langsmith https://twitter.com/LangChain/status/2077076144248021236
Many users praised LangSmith's Codex tracing plugin and agent support for enabling production-grade debugging and observability across tools like Cursor and Copilot, while one dismissed it as inferior to a competitor.
Based on 5 visible X reactions from 8 accounts; directional sample.
Ask a question below.
Published answers will appear here.
we did a bunch of work to standardize how coding agents trace to langsmith https://twitter.com/LangChain/status/2077076144248021236