Introduced affinity score, the value to evaluate the relevance of knowledges to given source. This enables us to write only on the topic that is considered familiar, with respect to the latent knowledges pre-acquired.
Improved ingestion pipeline that removes unnecessary content, including irrelevant comment.
Modified synthesis logic of latent-knowledges along with source documents.
May 20, 2025
Introduced ingestion pipeline, which enables any source documents to be cleaned and structured before persisted.
May 18, 2025
Introduced style retriever, including RAG-Similarity-Retriever, which retrieves documents from the pre-digested knowledge sources as a style documents upon the input source.
May 12, 2025
Introduced batch scheduler: automated ingestion, cleaning, synthesis, generation and posting are now possible.
May 11, 2025
New feature: additional tag option can be selected when posting.
May 8, 2025
Completed the first version of source retrieval workflow, involved status management along with distinguished meta data and the contents.