We've redesigned the LangSmith product homepage and made enhancements to Resource Tags for improved organization within your workspaces. The homepage is now organized into three key areas to align with core developer workflows: Observability, Evaluation, and Prompt Engineering. These three areas are broken down into the following:

• Observability (to identify issues quickly and understand your LLM app's performance over time)

  • Tracing Projects: Debug applications with datapoint-level trace analysis.

  • Dashboards: Monitor metrics like cost, latency, quality, and custom KPIs.

• Evaluation (to ensure continuous LLM app improvement by evaluating datasets and highlighting regressions)

  • Datasets: Curate and import data to fuel experiments.

  • Experiments: Run evaluations and track metrics like accuracy and style.

  • Annotation Queues: Streamline human reviews for better feedback.

• Prompt Engineering (to iterate and collaborate on prompts for optimization)

  • Prompts: Store and version prompts for team collaboration.

  • Playground: Iterate on prompts in real-time, test at scale, and compare results side-by-side.

LangSmith's Resource Tags now also offer greater flexibility for organizing workspace assets, including:

• Customizable Grouping: Tag resources by project, environment (e.g., dev, staging, prod), or custom categories.

• Improved Filtering: Quickly locate resources using default or custom tags.

• Coming Soon: Attribute-Based Access Control (ABAC) for fine-grained permissions based on tag attributes.

How to Get Started

1. Visit smith.langchain.com to explore the new homepage.

2. Use the Onboarding Guide to navigate Observability, Evaluation, and Prompt Engineering.

3. Experiment with Resource Tags to group and organize workspace assets.

4. Share your feedback using the in-app form—we’d love to hear your thoughts!