Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Search NIH-funded grants and research projects via RePORTER API
.claude/skills/brycewang-stanford-nih-reporter-api-guide/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 92% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 134% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 72% | 0% |
NIH RePORTER (Research Portfolio Online Reporting Tools) provides a comprehensive API for searching and analyzing grants funded by the National Institutes of Health and other agencies within the U.S. Department of Health and Human Services. The database includes project details, funding amounts, publications, patents, and clinical studies linked to funded research.
RePORTER is the authoritative source for NIH grant data and covers billions of dollars in annual biomedical research funding. The API enables programmatic access to search for funded projects, retrieve associated publications, and analyze funding trends across NIH institutes, study sections, and disease categories.
The v2 API is a modern RESTful service that accepts JSON POST bodies for search requests and returns structured JSON responses. It is completely free and requires no authentication.
No authentication is required. The NIH RePORTER API is a free public service.
bash# No API key needed curl -X POST "https://api.reporter.nih.gov/v2/projects/search" \ -H "Content-Type: application/json" \ -d '{"criteria":{"advanced_text_search":{"search_field":"terms","search_text":"CRISPR"}},"limit":5}'
The primary endpoint for finding NIH-funded grants and projects.
POST https://api.reporter.nih.gov/v2/projects/searchRequest body (JSON):
criteria.advanced_text_search: Text query with search_field and search_textcriteria.fiscal_years: Array of fiscal years (e.g., 2023, 2024, 2025])criteria.pi_names: Array of PI name objects with first_name, last_namecriteria.org_names: Array of institution namescriteria.agencies: Array of agency codes (e.g., "NIH"])criteria.activity_codes: Grant mechanism codes (e.g., "R01", "R21"])limit: Results per page (max 500)offset: Pagination offsetExample: Search for CRISPR gene editing R01 grants:
bashcurl -s -X POST "https://api.reporter.nih.gov/v2/projects/search" \ -H "Content-Type: application/json" \ -d '{ "criteria": { "advanced_text_search": { "search_field": "projecttitle,terms", "search_text": "CRISPR gene editing" }, "activity_codes": ["R01"], "fiscal_years": [2024, 2025] }, "limit": 10, "offset": 0 }' | python3 -m json.tool
Find publications linked to NIH-funded projects.
POST https://api.reporter.nih.gov/v2/publications/searchbashcurl -s -X POST "https://api.reporter.nih.gov/v2/publications/search" \ -H "Content-Type: application/json" \ -d '{ "criteria": { "core_project_nums": ["R01GM123456"] }, "limit": 25, "offset": 0 }' | python3 -m json.tool
pythonimport requests API_URL = "https://api.reporter.nih.gov/v2/projects/search" def search_nih_projects(query, fiscal_years=None, activity_codes=None, limit=50): """Search NIH RePORTER for funded projects.""" payload = { "criteria": { "advanced_text_search": { "search_field": "projecttitle,terms", "search_text": query } }, "limit": limit, "offset": 0 } if fiscal_years: payload["criteria"]["fiscal_years"] = fiscal_years if activity_codes: payload["criteria"]["activity_codes"] = activity_codes resp = requests.post(API_URL, json=payload) resp.raise_for_status() data = resp.json() return data.get("results", []), data.get("meta", {}).get("total", 0) results, total = search_nih_projects( "Alzheimer disease biomarkers", fiscal_years=[2024, 2025], activity_codes=["R01", "R21", "U01"] ) print(f"Total matching projects: {total}") institute_totals = {} for project in results: ic = project.get("agency_ic_fundings", []) for funding in ic: name = funding.get("abbreviation", "Unknown") amount = funding.get("total_cost", 0) or 0 institute_totals[name] = institute_totals.get(name, 0) + amount for inst, total_amt in sorted(institute_totals.items(), key=lambda x: -x[1]): print(f" {inst}: ${total_amt:,.0f}")
Grant Prospecting: Search for recently funded projects in your area to understand current NIH priorities, typical award sizes by mechanism (R01, R21, K99, etc.), and successful project framing.
Publication-Grant Linkage: Use the publications endpoint to find papers produced from specific grants, enabling analysis of research output and impact per dollar invested.
PI Network Analysis: Search by PI name to map a researcher's full NIH funding history, co-investigators, and institutional affiliations over time.
Funding Trend Tracking: Query across multiple fiscal years with consistent keywords to track how NIH investment evolves in emerging areas such as AI in healthcare, mRNA therapeutics, or long COVID.
offset for pagination)include_fields in the request body to limit response fields for faster responsesfiscal_years to narrow results and improve performance| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 17,406 | 29,454 | +69% | 1 | 1 | 0% | 3,472 | 5,148 | +48% | 0 | 0 | — |
case-02 | pass→pass | 10,054 | 4,842 | -52% | 1 | 1 | 0% | 1,616 | 2,391 | +48% | 0 | 0 | — |
case-03 | pass→pass | 9,407 | 5,864 | -38% | 1 | 1 | 0% | 1,718 | 2,594 | +51% | 0 | 0 | — |
case-04 | pass→pass | 11,377 | 5,197 | -54% | 1 | 1 | 0% | 1,805 | 2,520 | +40% | 0 | 0 | — |
case-05 | pass→pass | 8,557 | 3,231 | -62% | 1 | 1 | 0% | 1,075 | 2,276 | +112% | 0 | 0 | — |
case-06 | pass→pass | 7,604 | 3,090 | -59% | 1 | 1 | 0% | 1,135 | 2,195 | +93% | 0 | 0 | — |
case-07 | pass→pass | 8,105 | 4,579 | -44% | 1 | 1 | 0% | 1,341 | 2,740 | +104% | 0 | 0 | — |
case-08 | pass→pass | 10,732 | 10,426 | -3% | 1 | 1 | 0% | 2,165 | 3,350 | +55% | 0 | 0 | — |
case-09 | pass→pass | 10,757 | 7,628 | -29% | 1 | 1 | 0% | 1,929 | 2,838 | +47% | 0 | 0 | — |
case-10 | fail→pass | 11,934 | 7,189 | -40% | 1 | 1 | 0% | 2,129 | 2,772 | +30% | 0 | 0 | — |
case-11 | pass→pass | 9,981 | 6,042 | -39% | 1 | 1 | 0% | 1,472 | 2,602 | +77% | 0 | 0 | — |
case-12 | pass→pass | 11,661 | 7,201 | -38% | 1 | 1 | 0% | 1,811 | 2,799 | +55% | 0 | 0 | — |
case-13 | pass→pass | 7,304 | 6,168 | -16% | 1 | 1 | 0% | 1,375 | 2,769 | +101% | 0 | 0 | — |
case-14 | fail→pass | 17,585 | 5,468 | -69% | 1 | 1 | 0% | 3,135 | 2,711 | -14% | 0 | 0 | — |
case-15 | pass→pass | 11,485 | 9,224 | -20% | 1 | 1 | 0% | 2,375 | 3,625 | +53% | 0 | 0 | — |
case-16 | pass→pass | 10,713 | 11,419 | +7% | 1 | 1 | 0% | 2,292 | 3,509 | +53% | 0 | 0 | — |
case-17 | pass→pass | 8,725 | 2,680 | -69% | 1 | 1 | 0% | 1,490 | 2,070 | +39% | 0 | 0 | — |
case-18 | fail→pass | 7,528 | 4,870 | -35% | 1 | 1 | 0% | 1,252 | 2,403 | +92% | 0 | 0 | — |
case-19 | pass→pass | 3,683 | 1,963 | -47% | 1 | 1 | 0% | 470 | 1,898 | +304% | 0 | 0 | — |
case-20 | pass→pass | 9,511 | 3,082 | -68% | 1 | 1 | 0% | 1,582 | 2,278 | +44% | 0 | 0 | — |
case-21 | fail→pass | 4,770 | 2,078 | -56% | 1 | 1 | 0% | 799 | 1,870 | +134% | 0 | 0 | — |
case-22 | fail→pass | 16,397 | 11,642 | -29% | 1 | 1 | 0% | 2,373 | 4,078 | +72% | 0 | 0 | — |
case-23 | pass→pass | 13,202 | 10,470 | -21% | 1 | 1 | 0% | 2,355 | 3,511 | +49% | 0 | 0 | — |
case-24 | pass→pass | 15,067 | 13,086 | -13% | 1 | 1 | 0% | 3,152 | 4,318 | +37% | 0 | 0 | — |
case-25 | pass→pass | 18,742 | 14,247 | -24% | 1 | 1 | 0% | 3,076 | 4,293 | +40% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted. The headline lift of +20 percentage points is the difference between those two pass rates over the 25 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.