---
title: Research Papers Federated Search
emoji: 📚
colorFrom: indigo
colorTo: blue
sdk: gradio
sdk_version: 5.50.0
app_file: app.py
pinned: true
license: mit
mcp_server: true
tags:
- research
- papers
- academic-search
- federated-search
- mcp
- mcp-server
- arxiv
- pubmed
- semantic-scholar
- openalex
- scholarly
datasets:
- barissozudogru/openalex-concepts
- barissozudogru/federated-search-traces
---

# Research Papers Federated Search

## What this does

Federated search across four academic sources in a single call:

- **arXiv** -- Computer science, physics, math, quantitative biology
- **PubMed** -- Biomedical literature via NCBI E-utilities
- **Semantic Scholar** -- Cross-disciplinary, includes citation counts
- **OpenAlex** -- ~250M works with canonical IDs and concept tags

Results are deduplicated by DOI when available, then by `(source, source_id)` so the same arXiv preprint surfaced through multiple sources collapses to one row.

## Cold start

This Space sleeps when idle. First query after a quiet period takes 30-60 seconds for the container to wake up. Subsequent queries are fast.

## The full server

This Space demonstrates the core federated search. The complete MCP server has citation graphs, BibTeX export, persistent SQLite cache, topic detection with embeddings, and trend analysis:

```
pip install git+https://github.com/barissozudogru/deep-research-digest.git
```

## Related

- Source code: <https://github.com/barissozudogru/deep-research-digest>
- Maintainer: <https://bsozudogru.com>