Last updated
Graphlit now supports formatting of LLM-ready prompts with our RAG pipeline, via the new formatConversation and completeConversation mutations. This is valuable for supporting LLM streaming by directly calling the LLM from your application, and using Graphlit for RAG retrieval and conversation history. (Colab Notebook Example)
We have added support for inline hyperlinks in extracted text from documents and web pages.
GPLA-3466: Owner ID should accept any non-whitespace string
GPLA-3458: Not getting Person-to-Organization edges from entity extraction
