Has anyone tried integrating Scholarcy outputs with Rayyan for screening?

Hey everyone! I've been deep in the weeds of a systematic review project lately, and it's got me thinking about tooling pipelines. As many of you know, I'm always looking for ways to streamline data flow between systems—old habits from my database migration days die hard! 😅

I've been using Scholarcy for a while now to digest and summarize the mountains of PDFs I gather in the initial stages. Its ability to pull out key claims, methods, and results into a structured summary is a huge time-saver. Simultaneously, for the screening phase, I've been really impressed with Rayyan's collaborative screening and deduplication features. It's fantastic for quickly sorting the relevant from the irrelevant.

But here's my current friction point: it feels like I'm constantly context-switching and manually transferring data. My process looks something like this:

1. Scholarcy digests a batch of PDFs, giving me its lovely structured highlights.
2. I then have to go into Rayyan, often *re-uploading those same PDFs*, and manually re-tag or add notes based on Scholarcy's insights.

This feels inefficient and error-prone. I'm essentially maintaining two separate datasets. So, my big question is: **Has anyone successfully built a bridge between these two tools?**

I'm envisioning some kind of automated flow. Scholarcy can export its summary data (in JSON or even a formatted Markdown). Rayyan allows for importing studies via CSV and has API capabilities for adding notes/comments. Theoretically, one could write a script to parse Scholarcy's output and push the key findings as notes or custom fields into the corresponding study in Rayyan.

My initial thoughts for a potential integration pipeline would involve:
* Using Scholarcy's API (if available for your plan) or their exported JSON to get structured data.
* Mapping key Scholarcy fields (like 'Key Claims', 'Study Participants', 'Main Results') to Rayyan's note system or custom labels.
* Using Rayyan's API to find the matching study (likely by a consistent identifier like DOI) and attach the data.

A super basic conceptual snippet of what the transformation logic might look like (in pseudo-code):

```json
// Hypothetical Scholarcy output snippet for a single article
{
"doi": "10.1234/example.doi",
"key_claims": ["Claim A", "Claim B"],
"participants": "200 randomized patients",
"summary": "A brief automated summary text."
}
```

```python
# Hypothetical script step to format for Rayyan
# This would map to Rayyan's CSV import 'notes' field or via API
rayyan_note = f"""
SCHOLARCY EXTRACT:
Key Claims: {', '.join(scholarcy_data['key_claims'])}
Participants: {scholarcy_data['participants']}
Summary: {scholarcy_data['summary']}
"""
```

Has anyone tried something like this? I'm particularly curious about:
* **Practical Pitfalls:** Did you run into issues with data mapping, identifier matching, or API limits?
* **Workflow Impact:** Did this actually save time, or did it create a new layer of complexity?
* **Alternative Paths:** Is there a simpler middleman tool (like a well-structured Airtable or even a Python script with Pandas) that you used to glue them together?

I'm ready to roll up my sleeves and build a little connector if needed, but I'd love to learn from anyone who has already navigated these waters. Sharing your experiences (and code, if you have any!) would be invaluable.

—B