Step-by-step: Linking Scholarcy summaries to my Obsidian vault

I've been testing a workflow to get Scholarcy summary cards directly into my Obsidian vault for a few weeks now, and I think I've got a reliable pipeline. It's not a one-click solution, but it's pretty smooth once set up. The main goal was to have a clean, queryable database of paper summaries that I could link to my existing literature notes.

Here's my step-by-step approach, which leans heavily on the Scholarcy API and a bit of Python glue:

**1. The Core Setup: API to Markdown**
Scholarcy's API gives you a JSON output for any article you submit. I wrote a simple Python script that:
* Calls the API (using my API key).
* Parses the JSON to pull the key fields I care about: `title`, `abstract`, `key_concepts`, `summary_points`.
* Formats this into a clean Markdown file with YAML frontmatter for Obsidian.

```python
import requests
import yaml
import json

# Basic structure - you'd need to add your API key and handle the upload endpoint
def create_obsidian_note(api_response):
frontmatter = {
"title": api_response['title'],
"source": api_response['source_url'],
"authors": api_response['authors'],
"tags": ["#literature/summary"]
}
note_content = f"""---
{yaml.dump(frontmatter, default_flow_style=False)}
---
## Summary Points
{chr(10).join(['- ' + point for point in api_response['summary_points']])}

## Key Concepts
{chr(10).join(['- ' + concept for concept in api_response['key_concepts']])}

# Full abstract
{api_response['abstract']}
"""
return note_content
```

**2. The Automation Bridge**
I don't run this manually. I use a folder action on my Mac that watches the "Scholarcy Library" export folder. When a new `summary.json` file appears, the script triggers, creates the note, and moves it to my Obsidian vault's "Inbox" directory. On Windows, you could achieve this with Power Automate or a scheduled task.

**3. Linking & Enrichment in Obsidian**
Once the note is in the vault:
* I use the Dataview plugin to create a table of all summaries, which is super handy.
* I manually add links to related notes (e.g., `[[topic-machine-learning]]`) in the frontmatter after reading.
* Sometimes I'll append my own commentary in a separate section below the auto-generated content.

**Pitfalls & Considerations:**
* The API has rate limits, so batch processing a huge library takes some planning.
* The default summary points are great, but they're not tailored to my specific research questions. I still do a second pass to add my own highlights.
* You need a paid Scholarcy plan for API access – this isn't for the free tier.

Overall, it's turned Scholarcy from a standalone reading tool into a true input node for my PKM system. I'm curious if anyone else has tried this or uses a different method (Zapier? Readwise integration?). The main win for me is having all summaries in a format I can query with Dataview SQL-like queries.

--diver