Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Global biodiversity data API for species occurrences and datasets
.claude/skills/brycewang-stanford-gbif-api/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-24 | ✗→✓ | ▲ Improved | 118% | 0% |
| case-01 | ✓→✓ | = Same ✓ | 13% | 0% |
| case-02 | ✓→✓ | = Same ✓ | 56% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 26% | 0% |
The Global Biodiversity Information Facility (GBIF) is an international network and data infrastructure funded by governments worldwide, aimed at providing open access to biodiversity data. GBIF aggregates hundreds of millions of species occurrence records from natural history collections, citizen science platforms, monitoring networks, and published literature across the globe.
The GBIF API provides programmatic access to this vast repository of biodiversity data. Researchers can search for species occurrences by taxonomy, geography, time period, and dataset. The API also supports taxonomic name matching, dataset discovery, and species profile lookups. It serves as a foundational resource for ecological research, conservation planning, biogeography, and environmental impact assessments.
Ecologists, conservation biologists, biogeographers, and environmental scientists rely on the GBIF API to retrieve georeferenced occurrence data for species distribution modeling, climate change impact analysis, invasive species tracking, and biodiversity hotspot identification. The data is freely available under open data licenses.
No authentication required for read access. The GBIF API is publicly accessible without any API key or token. All search and retrieval endpoints are open. User authentication is only required for data publishing operations (creating datasets and uploading occurrences), which requires a GBIF account.
Search for georeferenced biodiversity observation and specimen records across all GBIF-indexed datasets.
GET https://api.gbif.org/v1/occurrence/search| Parameter | Type | Required | Description | |-----------------|--------|----------|------------------------------------------------------| | q | string | No | Full-text search query | | taxonKey | int | No | GBIF backbone taxonomy key | | scientificName | string | No | Scientific name to filter by | | country | string | No | ISO 3166-1 alpha-2 country code | | hasCoordinate | bool | No | Filter for georeferenced records only | | year | string | No | Year or range (e.g., 2020,2024) | | limit | int | No | Number of results (default 20, max 300) | | offset | int | No | Pagination offset |
bashcurl "https://api.gbif.org/v1/occurrence/search?scientificName=Panthera+tigris&hasCoordinate=true&limit=10"
count (total matches), results array with key, scientificName, decimalLatitude, decimalLongitude, country, basisOfRecord, eventDate, datasetKey, publishingOrgKey, and media links.Match a species name against the GBIF backbone taxonomy to resolve canonical names and get taxonomy keys.
GET https://api.gbif.org/v1/species/match| Parameter | Type | Required | Description | |-----------|--------|----------|------------------------------------------| | name | string | Yes | Scientific name to match | | kingdom | string | No | Kingdom filter for disambiguation | | strict | bool | No | If true, only return exact matches |
bashcurl "https://api.gbif.org/v1/species/match?name=Homo+sapiens"
usageKey, scientificName, canonicalName, rank, status, kingdom, phylum, class, order, family, genus, species, confidence score, and matchType.Search for and retrieve metadata about GBIF-indexed datasets from publishers worldwide.
GET https://api.gbif.org/v1/dataset| Parameter | Type | Required | Description | |-----------------|--------|----------|-------------------------------------------------| | q | string | No | Full-text search query | | type | string | No | Dataset type: OCCURRENCE, CHECKLIST, etc. | | publishingOrg | string | No | Publishing organization UUID | | limit | int | No | Number of results (default 20, max 1000) | | offset | int | No | Pagination offset |
bashcurl "https://api.gbif.org/v1/dataset?q=bird+monitoring&type=OCCURRENCE&limit=5"
count, results array with key, title, description, type, publishingOrganizationKey, license, recordCount, and endpoints.No formal rate limits are enforced on the GBIF API. However, GBIF recommends responsible use patterns. Large data downloads (millions of records) should use the asynchronous download API at https://api.gbif.org/v1/occurrence/download/request rather than paginating through the search endpoint. The search endpoint is limited to 100,000 records maximum per query via pagination.
Retrieve georeferenced occurrence data for species distribution modeling:
pythonimport requests params = { "taxonKey": 2480498, # Panthera tigris "hasCoordinate": True, "limit": 300 } resp = requests.get("https://api.gbif.org/v1/occurrence/search", params=params) data = resp.json() coordinates = [(r["decimalLongitude"], r["decimalLatitude"]) for r in data["results"] if "decimalLongitude" in r and "decimalLatitude" in r] print(f"Retrieved {len(coordinates)} georeferenced occurrences of {data['results'][0]['scientificName']}")
Resolve a list of species names against the GBIF backbone taxonomy:
pythonimport requests names = ["Homo sapiens", "Canis lupus", "Quercus robur", "Drosophila melanogaster"] for name in names: resp = requests.get("https://api.gbif.org/v1/species/match", params={"name": name}) match = resp.json() print(f"{name} -> {match['canonicalName']} (key: {match['usageKey']}, confidence: {match['confidence']})")
For large-scale analyses requiring millions of records, use the asynchronous download API:
bashcurl -X POST "https://api.gbif.org/v1/occurrence/download/request" \ -H "Content-Type: application/json" \ -u username:password \ -d '{"creator":"username","predicate":{"type":"equals","key":"TAXON_KEY","value":"2480498"}}'
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 15,682 | 24,037 | +53% | 1 | 1 | 0% | 3,281 | 3,692 | +13% | 0 | 0 | — |
case-02 | pass→pass | 11,545 | 7,823 | -32% | 1 | 1 | 0% | 2,201 | 3,423 | +56% | 0 | 0 | — |
case-03 | pass→pass | 18,035 | 14,026 | -22% | 1 | 1 | 0% | 3,571 | 4,492 | +26% | 0 | 0 | — |
case-04 | pass→pass | 6,128 | 6,088 | -1% | 1 | 1 | 0% | 1,220 | 2,702 | +121% | 0 | 0 | — |
case-05 | pass→pass | 10,349 | 4,020 | -61% | 1 | 1 | 0% | 1,695 | 2,484 | +47% | 0 | 0 | — |
case-06 | pass→pass | 15,507 | 6,334 | -59% | 1 | 1 | 0% | 2,544 | 3,091 | +22% | 0 | 0 | — |
case-07 | pass→pass | 10,834 | 4,055 | -63% | 1 | 1 | 0% | 1,711 | 2,395 | +40% | 0 | 0 | — |
case-08 | fail→pass | 16,554 | 12,590 | -24% | 1 | 1 | 0% | 3,103 | 4,386 | +41% | 0 | 0 | — |
case-09 | pass→pass | 16,335 | 21,051 | +29% | 1 | 1 | 0% | 2,751 | 5,250 | +91% | 0 | 0 | — |
case-10 | pass→pass | 14,658 | 6,142 | -58% | 1 | 1 | 0% | 2,416 | 3,095 | +28% | 0 | 0 | — |
case-11 | pass→pass | 9,384 | 5,085 | -46% | 1 | 1 | 0% | 1,192 | 2,484 | +108% | 0 | 0 | — |
case-12 | pass→pass | 12,667 | 11,165 | -12% | 1 | 1 | 0% | 2,114 | 3,567 | +69% | 0 | 0 | — |
case-13 | pass→pass | 8,369 | 8,272 | -1% | 1 | 1 | 0% | 1,361 | 3,002 | +121% | 0 | 0 | — |
case-14 | pass→pass | 11,739 | 13,834 | +18% | 1 | 1 | 0% | 2,395 | 4,053 | +69% | 0 | 0 | — |
case-15 | pass→pass | 13,573 | 12,625 | -7% | 1 | 1 | 0% | 2,279 | 3,812 | +67% | 0 | 0 | — |
case-16 | pass→pass | 18,633 | 11,477 | -38% | 1 | 1 | 0% | 3,143 | 4,080 | +30% | 0 | 0 | — |
case-17 | pass→pass | 19,197 | 10,645 | -45% | 1 | 1 | 0% | 2,921 | 3,750 | +28% | 0 | 0 | — |
case-18 | pass→pass | 13,120 | 10,661 | -19% | 1 | 1 | 0% | 1,983 | 4,004 | +102% | 0 | 0 | — |
case-19 | pass→pass | 5,261 | 1,732 | -67% | 1 | 1 | 0% | 1,136 | 2,016 | +77% | 0 | 0 | — |
case-20 | pass→pass | 13,624 | 6,000 | -56% | 1 | 1 | 0% | 2,187 | 2,933 | +34% | 0 | 0 | — |
case-21 | pass→pass | 22,806 | 10,794 | -53% | 1 | 1 | 0% | 2,366 | 3,360 | +42% | 0 | 0 | — |
case-22 | pass→pass | 16,009 | 11,754 | -27% | 1 | 1 | 0% | 3,356 | 4,013 | +20% | 0 | 0 | — |
case-23 | pass→pass | 10,653 | 5,181 | -51% | 1 | 1 | 0% | 1,762 | 2,791 | +58% | 0 | 0 | — |
case-24 | fail→pass | 6,738 | 5,183 | -23% | 1 | 1 | 0% | 1,159 | 2,522 | +118% | 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. 24 cases were attempted. The headline lift of +8 percentage points is the difference between those two pass rates over the 24 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.