How to enhance Knowledge Retrieval Node Input for Improved Accuracy

[Spirit597]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

Content

I'm facing issues with the Knowledge Retrieval node in Dify workflows.

Directly using the user's query as input often leads to irrelevant results, especially with mixed correct/incorrect terms. This means we lose critical knowledge base information in chat assistants and Chatflow.

Alternatively, extracting comma-separated keywords also falls short. The order and continuity of these keywords significantly impact retrieval, potentially missing crucial context.

I believe optimizing the input method would greatly boost the node's effectiveness. Have you considered or are there plans for:

• Smarter query parsing for better natural language and noisy inputs?
• Intelligent keyword extraction/embedding-based search considering semantic similarity?
• Leveraging the original query's structure and word relationships more effectively?