March 12, 2025

Overview

In this release, we’ve added the following features:

• Images support for Direct Mode in stateless endpoint
• Custom icons for knowledge models branding

What's new

Images support for Direct Mode in stateless endpoint

The stateless endpoint now supports image inputs for multimodal language models (LLMs) in direct mode. This enhancement enables more dynamic interactions by using visual data in processing.

Image support is fully compatible with popular frameworks like LangChain and OpenAI, ensuring seamless integration and efficient use of both image and text data in the API.

Custom icons for knowledge models branding

Knowledge model owners can now set a custom logo or icon to personalize their models. This feature enhances branding by prominently displaying the logo in chats with the knowledge model, reinforcing its visual identity and recognition.