LangSmith now supports images, PDFs, and audio files across the playground, annotation queues, and datasets — making it easier than ever to build, test, and evaluate multimodal applications.

LangSmith helps you bring real-world data into your workflows. You can now:

• Attach files directly to dataset examples, eliminating the need for clunky base64 encoding

• Speed up evaluations with faster upload/download of binary files

• Visualize multimodal content right inside the LangSmith UI for smoother debugging and iteration

Watch it in action:

How we evaluate a receipt-extracting agent

Ready to build?

Upload images, PDFs, or audio to your datasets and run your first multimodal eval:

• Get started with multimodal evaluation in the docs

• Try it live in the LangSmith playground