β π LangGraph Cloud: Deploy at scale, monitor carefully, iterate boldly
LangGraph Cloud is now in closed beta, offering scalable, fault-tolerant deployment for LangGraph agents. Deploy in one-click and get an integrated tracing &...
π Self-improving LLM evaluators in LangSmith
Using an β LLM-as-a-Judge β is a popular way to grade outputs from LLM applications. This involves passing the generated output to a separate LLM and asking...
βοΈ Trim messages with transformer utils
Processing LangChain messages is now easier in both Python and JavaScript with our new message transformers. The new trim_messages util in LangChain covers a...
π§· Store the model and configuration when saving a prompt
When you save a prompt in LangSmith, the model and its configuration will also be stored. You no longer need to reset your preferences in the Playground...
π Custom models in LangSmith Playground
Once youβve deployed a model server, use it in the LangSmith Playground by selecting either ChatCustomModel or the CustomModel provider for chat-style model...
π₯· PII Masking in LangSmith
LangSmith SDK versions now have enhanced PII masking capabilities. You can easily create anonymizers by specifying a list of regular expressions or providing...
π Workspaces in LangSmith for improved collaboration & organization
LangSmith now offers Workspaces to separate resources for different teams or different environments. Admins can add users to Workspaces, granting them...
πΊοΈ Variable mapping for online evaluator prompts
In LangSmith, you can now use any structured prompt from the LangChain Hub with customizable variables. You can personalize your inputs based on recent runs...
π Enter the playground from scratch instead of from a trace or a prompt
The Playground is now its own tab in the sidebar of LangSmith. To create a new prompt , simply craft a prompt in the empty playground and click "Save As" to...
π¬ Chat LangChain improvements
Users can now view and continue previous chats in Chat LangChain , our chatbot for LangChain Python docs and API reference. These improvements were made...