LangSmith Custom Apps: Build custom interfaces around your agent data

LangChain on September 24, 2026 launched Custom Apps in LangSmith, a feature that lets teams build, publish, and run custom user interfaces on top of LangSmith data directly inside LangSmith. The company said the launch skips the hosting, authentication, permissions, and sharing work that custom frontends typically require.
Teams can start with a chat prompt or build in code. On Plus and Enterprise plans, LangSmith Chat can turn a description of the desired data and visualization into a working app. Alternatively, users can start from provided templates and build against the LangSmith API with a coding agent. Once published, the app becomes a shared interface inside the workspace for recurring workflows, rather than exporting data or rebuilding static views.
LangChain described common uses: annotation interfaces that give reviewers context matching their workflows, such as full traces for engineers or a rubric for subject matter experts; experiment review UIs that compare outputs across prompt versions, model swaps, customer segments, or failure categories; and focused trace review views built around frequent questions.
Custom Apps use Macaw, LangSmith's design system, by default, and the company said it is open sourcing Macaw. Users can customize the look via chat or a coding agent.
Chime's Roxanne Wang said in a provided quote that human-in-the-loop reviews previously lived across spreadsheets and separate tools, and the team built a Custom App for structured LLM evaluation and annotation that runs inside LangSmith and connects through the API, allowing review queues and dashboards.
Availability is limited to Plus and Enterprise plans. Plus includes one Custom App per organization and the chat experience; Enterprise includes unlimited Custom Apps with the chat experience.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
LangSmith Custom Apps lets you build the interface you want with your LangSmith data, publish it into your workspace, and skip the hosting, auth, and permissions work. Learn more.