Langchain
Day 19 after signup
LangSmith Engine: The agent for agent engineering
Once traces are flowing, Engine improves your agents autonomously.
Why it works
- Open with the problem the product addresses before explaining its capabilities, so readers have context for the announcement.
- Break a complex feature into clearly named functions, then provide direct paths to documentation and pricing.
- Keep the main explanation concise and place support and learning resources after the primary product links.
How it is built
- Product name header
- Agent engineering headline
- Problem setup
- Capability breakdown
Email copy
LANGSMITH ENGINE The agent for agent engineering When agents are running in production, manual trace review only gets you so far. LangSmith Engine helps. Available on Plus, Startup, and Enterprise plans, Engine drives fixes across the entire agent development lifecycle: Issue detection: Monitors production traces for evaluator failures, tool errors, latency spikes, and negative user feedback. Related failures are grouped into named issues with severity, timeline, and trace evidence. Root cause diagnosis: When your repo is connected, Engine reads the relevant code, identifies the likely issue, and opens a GitHub PR with the proposed fix for your review. Eval coverage: Proposes an online evaluator scoped to the failure pattern and adds failing traces to your offline eval dataset, so the same issue can't silently recur. Read docs → View pricing → Need help? 💬 Support Portal: Find answers, submit a ticket, or chat with our AI assistant. 📙Read the LangSmith Engine announcement blog. 🎥 Academy: Self-paced courses on agent observability, evals, deployment, and more. LangChain, 2261 Market Street, STE 22961, San Francisco, California 94114, United States Unsubscribe Manage preferences