---
title: "Sugra Research"
description: "Scholarly and development data: arXiv, NBER, Crossref, and long-run growth statistics."
slug: sugra-research
status: published
published_at: 2026-06-30
author: Arman Obosyan
author_url: https://sugra.systems/about
section: foundation
series_order: 7
primary_keyword: sugra research arxiv api
hero_image: /blog/images/posts/sugra-research-hero.jpg
hero_alt: "Sugra Research - Sugra API blog"
og_image: /blog/images/posts/sugra-research-hero.jpg
tags:
  - research
  - arxiv
  - api
---

# Sugra Research: preprints, scholarly metadata, and long-run development data

**Series:** Sugra product directions (7 of 7)  
**Audience:** developers integrating the Sugra API over HTTP  
**Live samples verified:** 2026-07-24 against `https://sugra.ai`  
**Auth:** `x-api-key` header on every data call  

Standalone article. Sugra Research covers the scholarly record and long-run development statistics - from this morning's preprints to multi-century growth panels.

---

## Essence

> Scholarly and development data - from this morning's preprints to a century of growth statistics.

Core surfaces:

- arXiv preprint search, paper fetch, categories  
- Crossref DOI works, journals, funders, references, citations  
- NBER working papers and (when available) recession dates  
- Maddison Project long-run GDP and population  
- Penn World Table growth and productivity panel  
- Adjacent development slices: WHO GHO, OWID composites, UN/ILO/FAO series (often under neighboring reference categories but one key)

All academic or intergovernmental sources, named in `meta.source` (for example `arxiv`, `crossref`, `maddison`, `pwt`, `nber`).

---

## What this direction is not

| Not this | Why |
|---|---|
| Full-text paper hosting | Metadata, links, and abstracts - not a pirate library |
| Citation manager product | Graph walks via API, you store the library |
| Official high-frequency macro prints | **Sugra Macro** (FRED, BIS, Treasury) |
| News about research | **Sugra News** |
| Investment advice from papers | Literature access only |

---

## Capability map

| Category | Endpoints (approx.) | Coverage |
|---|---|---|
| Research core | 34 | arXiv (7), Crossref (10), NBER (2), Maddison (7), PWT (8) under `/api/v1/research/*` |
| Health | 10 | WHO Global Health Observatory |
| Adjacent development | varies | OWID, UN Population, ILOSTAT, SDG, FAOSTAT live in related categories with the same key |

---

## Why the response shape is useful

### 1. Fielded arXiv search with full abstract and URLs

Live search `q=agentic+reasoning` (limit 2): **680** total hits; papers include `arxiv_id`, title, full summary, authors, categories, published timestamps, `urls.pdf` / `urls.abstract`.

```json
{
  "data": {
    "query": "all:\"agentic reasoning\"",
    "papers": [
      {
        "arxiv_id": "2601.12538",
        "title": "Agentic Reasoning for Large Language Models",
        "primary_category": "cs.AI",
        "published": "2026-01-18T18:58:23Z",
        "urls": {
          "abstract": "https://arxiv.org/abs/2601.12538v1",
          "pdf": "https://arxiv.org/pdf/2601.12538v1"
        }
      }
    ],
    "pagination": { "total_results": 680, "limit": 2 }
  },
  "meta": { "source": "arxiv" }
}
```

### 2. Crossref DOI resolution with citation graph hooks

Live DOI `10.1038/s41586-021-03819-2` (AlphaFold paper): title, venue Nature, authors, license, `is-referenced-by-count` (43000+), references list, publisher metadata. Sibling routes: references and citations by DOI.

### 3. Long-run Maddison panels

Country series with year, GDP per capita, population - centuries of observations for many countries. Methodology and citation live under definitions endpoints.

### 4. PWT rankings and TFP panels

Rank countries by real GDP or productivity variables with attribution strings suitable for academic footnotes.

### 5. NBER latest papers

`/api/v1/research/nber/papers/latest` - live 200. Recession dates endpoint can 503 when static fetch fails - handle in clients.

---

## Developer path

```bash
export SUGRA_API_KEY="YOUR_API_KEY"

curl -H "x-api-key: $SUGRA_API_KEY" \
  "https://sugra.ai/api/v1/research/arxiv/search?q=agentic+reasoning&limit=5"

curl -H "x-api-key: $SUGRA_API_KEY" \
  "https://sugra.ai/api/v1/research/crossref/works/10.1038/s41586-021-03819-2"

curl -H "x-api-key: $SUGRA_API_KEY" \
  "https://sugra.ai/api/v1/research/maddison/country/JPN"

curl -H "x-api-key: $SUGRA_API_KEY" \
  "https://sugra.ai/api/v1/research/nber/papers/latest?limit=5"
```

```python
import os, requests
H = {"x-api-key": os.environ["SUGRA_API_KEY"]}
r = requests.get(
    "https://sugra.ai/api/v1/research/arxiv/search",
    headers=H, params={"q": "climate", "limit": 10}, timeout=30,
)
r.raise_for_status()
papers = r.json()["data"]["papers"]
```

---

## Use case 1: Morning preprint scan

**Goal.** What appeared on arXiv for a topic or category.

**Calls.** `/research/arxiv/search`, `/research/arxiv/latest?cat=`, `/research/arxiv/paper/{id}`.

**Strength.** Structured papers + pdf links + pagination.  
**Pitfalls.** Query syntax is arXiv-style; relevance sort is not peer review.

---

## Use case 2: DOI resolve and cite

**Goal.** Turn a DOI into authoritative metadata for an agent or UI.

**Call.** `/research/crossref/works/{doi}`.

**Strength.** Venue, authors, licenses, citation counts.  
**Pitfalls.** Huge payloads - select fields client-side for LLM context windows.

---

## Use case 3: Citation graph walk

**Goal.** What a paper cites / who cites it.

**Calls.** Crossref references and citations routes for a DOI.

**Strength.** Enough for a lightweight graph without a full bibliometrics platform.  
**Pitfalls.** Coverage depends on Crossref deposits.

---

## Use case 4: Long-run GDP context

**Goal.** Place a country in multi-century GDP/pop history.

**Call.** `/research/maddison/country/{iso3}` (+ world/regional/top).

**Strength.** Teaching and research panels with methodology endpoints.  
**Pitfalls.** Historical estimates carry large uncertainty - show source attribution.

---

## Use case 5: Productivity and PWT rankings

**Goal.** Rank countries by real output or TFP for a year.

**Call.** `/research/pwt/top?variable=rgdpo&year=2019&limit=10`.

**Strength.** Standard growth-accounting panel with attribution.  
**Pitfalls.** Variable codes require the definitions endpoint.

---

## Use case 6: NBER working paper feed

**Goal.** Latest economics working papers for a research agent.

**Call.** `/research/nber/papers/latest`.

**Strength.** RSS-backed latest set with abstracts.  
**Pitfalls.** `pub_date` often null - order by paper id when recency matters.

---

## Use case 7: Health indicator scorecard (WHO)

**Goal.** Country health indicators and rankings.

**Calls.** `/api/v1/health/who/*` family (catalog, country data, latest, rankings).

**Strength.** Named WHO GHO series under the same key.  
**Pitfalls.** Health indicators lag; document as_of.

---

## Use case 8: Literature-grounded agent answer

**Goal.** Agent must not invent papers - search arXiv/Crossref, return titles + DOI/arXiv ids + `meta.source`.

**Pattern.**

1. Search arXiv or resolve DOI  
2. Quote title, year, venue  
3. Attach pdf/doi link  
4. Cite `meta.source`

**Strength.** Primary Research product for AI agent builders.  
**Pitfalls.** Always distinguish preprint from peer-reviewed venue.

---

## Antipatterns

1. Dumping entire Crossref JSON into an LLM prompt  
2. Presenting preprints as settled fact  
3. Omitting Maddison/PWT attribution  
4. Confusing Research long-run GDP with Macro FRED monthly CPI  
5. Hard-failing the product when NBER recession-dates returns 503 - degrade gracefully  

---

## Scenarios where Research alone is enough

- Academic search agents  
- DOI metadata services  
- Long-run development teaching tools  
- NBER watch feeds  
- WHO-based health scorecards  

For "today's CPI print" use Macro. For "stock reaction to a paper" that is a multi-direction composition (later series articles).

---

## Live verification (2026-07-24)

| Call | Result |
|---|---|
| arXiv search agentic reasoning | 200, 680 hits |
| Crossref AlphaFold DOI | 200 |
| Maddison JPN | 200 |
| NBER papers/latest | 200 |
| NBER recession-dates | 503 (handle as optional) |

---

## Where to go next

- Portal: Sugra Research direction + sources  
- Series complete for standalone directions  
- **Next phase (later):** composition articles that combine directions (for example Macro + Finance, Entity + News)  
- Cookbook and portal recipes for multi-call patterns  

| # | Direction | Status |
|---|---|---|
| 1 | Finance | Published |
| 2 | Macro | Published |
| 3 | Entity | Published |
| 4 | NetAtlas | Published |
| 5 | News | Published |
| 6 | Earth | Published |
| 7 | **Research** | This article |

---