What it does
A daily cron at 06:00 UTC iterates every active topic across all Pro users. For each topic it:
- Searches Europe PMC for open-access papers published since the topic's last run
- Drops anything without a downloadable PDF (closed-access papers are filtered out)
- Skips papers already linked to this vault (idempotent re-runs)
- Downloads each PDF through an SSRF-safe allowlist (europepmc.org, ncbi.nlm.nih.gov, mdpi.com, etc.)
- Runs the same ingestion pipeline as if you'd uploaded the PDF manually
- Stops when the per-topic daily cap is hit (default 5)
Creating a topic
From /settings → ScholarFlow → New topic. Fields:
- Query — Europe PMC search syntax (e.g.
"BPC-157 tendon repair") - Target vault — which vault the auto-ingested papers go into
- Daily ceiling — max papers per cron run (1-20; default 5)
- Paused — temporarily disable without deleting
How many topics, how often
| Plan | Topics | Auto-cron | Daily ingest cap |
|---|---|---|---|
| Free | — | No | 1 paper/day (manual only) |
| Pro | 5 | Yes — once daily | 100 papers/day across all topics |
Cost protection
The cron checks the org-wide daily cost cap ($50) before running. If yesterday's spend exceeded the cap, the whole pass is skipped and an audit-log entry is written. Per-Pro-user: if yourcost_used_today_usd is already over your cost_cap_day_usd ($5), your topics are skipped for that day.
What about closed-access papers?
Europe PMC's metadata covers most of the literature, but only open-access papers have a PDF we can index. Closed-access papers are visible in manual ScholarFlow search (Phase 8 feature) but the auto-cron never enqueues them. If you have institutional access to a closed-access paper, download it manually and add via the regular source upload.
Pausing or removing
Toggle Paused on a topic to skip it on the next cron without deleting it. Delete the topic permanently when you're done with that research thread. Existing papers stay in the vault when a topic is deleted; only the auto-update plumbing is removed.