Built a domain-specific research alert system on OpenAlex — AI + Materials Science tracker

[Indra sena reddy Katta]

Hi OpenAlex community,

I'm Indra, an AI Research Intern 

As part of our research workflow, I needed a reliable way to track new publications at the intersection of AI and Materials Science — without manual effort every day. So I built Crystal Research Intelligence on top of the OpenAlex API.

What it does:
- Queries OpenAlex daily using a strict AND query: (Material Keywords) AND (AI Keywords)
- Covers superconductors, solid-state batteries, rare earth materials, and AI-driven discovery
- Filters by publication date across a rolling 180-day window
- Deduplicates results using normalized DOI values stored in seen.json
- Persists metadata in report_data.json for historical accumulation
- Auto-generates and deploys a public searchable dashboard via GitHub Pages
- Sends instant alerts to Telegram and Discord on new papers
- Runs entirely on GitHub Actions — zero infrastructure cost

Live Dashboard: https://indra2215.github.io/research-monitor/
Repository: https://github.com/indra2215/research-monitor

I have two specific questions for the community:

1. Query strategy — I'm currently using a keyword AND approach. Are there better OpenAlex filter combinations (concepts, topics) that would improve precision for niche domain tracking like this?

2. Coverage gaps — Has anyone noticed OpenAlex missing recent papers in materials science or applied AI? I want to understand if my 50-result-per-run cap is the bottleneck or if it's an indexing lag.

The system is domain-agnostic by design. The same architecture can be repointed to drug discovery, climate tech, or quantum computing with only a config change — so I'm curious if others in this group have built similar domain-specific trackers and what challenges they ran into.

Any feedback appreciated.

Best,
Indra
https://github.com/indra2215/research-monitor