Skip to content

[Feature] Enhance webpage content parsing to match the quality of Obsidian Web Clipper/Sider #840

@ansewuLJ

Description

@ansewuLJ

The current webpage content extraction often results in disordered or incomplete text, which leads to chaotic or poor-quality AI responses.

Specific Issues:

  • The order of the extracted content is often mixed up.
  • Parts of the content are missing (incomplete parsing).

Comparison:
When using browser extensions like Obsidian Web Clipper or Sider, the resulting Markdown content is nearly perfect and well-structured. Page Assist currently falls short in this area, making the "chat with page" feature nearly unusable for complex websites.

Reproduction:
You can test this on the following URL to see the difference in parsing quality:
https://docx.dawclaudecode.com/

Expected Behavior:
Please optimize the parsing logic to ensure the extracted text maintains the correct order and completeness, similar to the quality provided by the tools mentioned above.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions