If using Obsidian's Web Clipper, the default or custom templates can have custom note sections which extract different combinations of data from webpages. Here is one example spun up by Claude.

Safaris with Obsidian.yaml

---
# Metadata header (these go in Properties section)
---

# {{title}}

## Metadata
- **Author:** {{author}}
- **Published:** {{published|date:"YYYY-MM-DD"}}
- **Captured:** {{time|date:"YYYY-MM-DD HH:mm"}}
- **Source:** {{url}}
- **Domain:** {{domain}}
- **Description:** {{description}}

## Highlights
{{highlights}}

## Main Content
{{content}}

## Full Page Text (Fallback)
{{text}}

## Raw HTML (if needed for technical pages)
{{contentHtml|markdown}}

## Page Metadata
- **Word Count:** {{wordCount}}
- **Reading Time:** {{readingTime}}
- **Image Count:** {{imageCount}}

## Schema.org Data (if available)
{{schema}}

## Tags
{{tags}}

---
## Notes
[Your analysis and notes here]

## Related
-

---
*Clipped from: [{{title}}]({{url}})*

Additional discoveries after getting this fired up:

• in settings for this extension you can create templates
• the templates can hold many variables as well as bracketed analysis instructions for interpreters
• numerous AI services can be connected into this extension and "Interpreter" can be run when the pop up has been activated for any viewed webpage
• one or several models can be loaded into the menu for each AI service
• once Interpreter runs the Note is repopulated with whatever your template determines
• save to Obsidian or to file!
• Properties section harnesses the best of Obsidian so tables can be viewed/sorted later with filtering
• create fields without corresponding variables from pages to manually add your own labeled inputs

Research Process Opportunity:

• let the different templates represent the different hats / roles you play for your company
• harness this to achieve / maintain the "corporation of 1"
• running Interpreter in the pop up ins't necessary, the formatting of the templated saved document along with the pre-populated prompts is enough to run a mass analysis by AI later and take work to next stages