LangChain Launches LangSmith Engine to Automatically Detect and Fix Agents

LangChain

Jun 15, 2026

LangChain launched LangSmith Engine, an agent that monitors production traces to identify and resolve agent failures. Engine clusters related issues, writes prompt or code fixes, and can open GitHub pull requests for review. It also automatically suggests online evaluators and offline datasets to prevent recurring errors, building a feedback loop from production failures to agent improvements.

View the full update on langchain.com

LangChain

@LangChain3d ago

LangSmith Engine goes beyond the addressing and diagnosis of problems, and actually works to fix them. A 1 minute explanation from @bentannyhill https://t.co/ae6v1oBqOC

540

View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Keep reading

LangSmith Engine Automates Agent Issue Resolution with PRs and Evals

LangChain's LangSmith Engine now automatically proposes three resolution actions for every agent issue it identifies: opening a Pull Request (PR), creating a custom online evaluator, and adding failing traces to an offline evaluation suite. This aims to accelerate the agent development lifecycle by automating issue diagnosis and fix validation.

ReplitApr 30

Replit Launches AI App Monitoring to Automate Production Incident Response

Replit launched App Monitoring, a production-focused toolset that enables its AI agent to autonomously diagnose and fix issues in live applications. The system moves beyond code generation to handle the Day 2 operations of software, allowing the agent to sift through production logs and databases to identify root causes.

Vercel Labs Launches Real-Time Dashboard to Debug AI Agent Browser Sessions

VercelMar 28

Vercel Labs Launches Real-Time Dashboard to Debug AI Agent Browser Sessions

Vercel Labs introduced a local web dashboard for its agent-browser CLI that provides a live visual feed and activity log of AI agent interactions. This observability layer addresses the black-box nature of headless automation by letting developers monitor browser console, network, and storage state in real time.

AWS Launches Bedrock AgentCore Optimization to Automate Agent Performance Loops

Amazon Web ServicesMay 5

AWS Launches Bedrock AgentCore Optimization to Automate Agent Performance Loops

AWS introduced AgentCore Optimization in preview to help developers systematically improve AI agent reliability through automated prompt and tool refinement. The platform analyzes real execution traces to generate optimized instructions and validates them using managed A/B experiments and batch testing.