Agent-Skills.md

Agent Skills: HathiTrust Bibliographic API Skill

Look up bibliographic records and digitized volumes in HathiTrust's 17M+ volume collection by ISBN, OCLC, LCCN, ISSN, or HathiTrust ID. Use this skill whenever the user wants to check if a book has been digitized, find which libraries hold a copy, get MARC metadata for a known identifier, or link a bibliographic reference to its HathiTrust digital version. Also triggers when cross-referencing identifiers from other library catalogs (Harvard, Library of Congress) against HathiTrust holdings.

UncategorizedID: kltng/humanities-skills/hathitrust-catalog

Install this agent skill to your local

pnpm dlx add-skill https://github.com/kltng/humanities-skills/tree/HEAD/hathitrust-catalog

Skill Files

Browse the full folder contents for hathitrust-catalog.

Download Skill

Loading file tree…

hathitrust-catalog/SKILL.md

Skill Metadata

Name
hathitrust-catalog
Description
Look up bibliographic records and digitized volumes in HathiTrust's 17M+ volume collection by ISBN, OCLC, LCCN, ISSN, or HathiTrust ID. Use this skill whenever the user wants to check if a book has been digitized, find which libraries hold a copy, get MARC metadata for a known identifier, or link a bibliographic reference to its HathiTrust digital version. Also triggers when cross-referencing identifiers from other library catalogs (Harvard, Library of Congress) against HathiTrust holdings.

HathiTrust Bibliographic API Skill

Look up records and digitized volumes in HathiTrust's collection of 17M+ volumes from major research libraries.

Critical: Things Claude Won't Know Without This Skill

This is a lookup API, NOT a search API

HathiTrust's Bibliographic API retrieves records by known identifiers only. There is no keyword or title search. You must have an ISBN, OCLC number, LCCN, ISSN, or HathiTrust ID.

WORKS:     https://catalog.hathitrust.org/api/volumes/brief/isbn/0140449264.json
NOT POSSIBLE: https://catalog.hathitrust.org/api/volumes/brief/search?q=hamlet

If you only have a title/author, search another catalog first (LibraryCloud, Library of Congress) to get an ISBN or OCLC number, then look it up here.

Two endpoint variants: brief vs full

| Endpoint | Returns | Use when | |----------|---------|----------| | /api/volumes/brief/ | Metadata + items (no MARC) | You just need titles, ISBNs, holdings, digital access links | | /api/volumes/full/ | Same + complete MARC-XML | You need full cataloging data |

Empty results are not errors

When no record matches, the API returns {"records": {}, "items": []} with HTTP 200 — not a 404. Always check if records is empty.

Batch up to 20 lookups in one request

Use the multi-ID endpoint to batch lookups:

/api/volumes/brief/json/isbn:0140449264|oclc:228668653

Pipe-separated as type:value. Max 20 per request.

Identifier Types

| Type | Example | Notes | |------|---------|-------| | isbn | 0140449264 or 9780140449266 | ISBN-10 or ISBN-13, digits only (+ trailing X) | | oclc | 228668653 | OCLC number, digits only | | lccn | 72081773 | URL-encode spaces/slashes | | issn | 03785955 | Digits only | | htid | mdp.39015058510069 | HathiTrust volume ID | | recordnumber | 000578050 | 9-digit HathiTrust record number |

Response Structure

{
  "records": {
    "009058846": {
      "recordURL": "https://catalog.hathitrust.org/Record/009058846",
      "titles": ["The Odyssey / Homer ; translated by Robert Fagles."],
      "isbns": ["0140268863", "9780140268867"],
      "oclcs": ["34228839"],
      "lccns": ["96017900"],
      "publishDates": ["1997"]
    }
  },
  "items": [
    {
      "orig": "University of Michigan",
      "htid": "mdp.39015042182054",
      "itemURL": "https://babel.hathitrust.org/cgi/pt?id=mdp.39015042182054",
      "rightsCode": "ic",
      "usRightsString": "Limited (search-only)",
      "lastUpdate": "20240115",
      "enumcron": false
    }
  ]
}

Key fields:

  • rightsCode: pd = public domain (full text available), ic = in-copyright (search only)
  • usRightsString: "Full View" or "Limited (search-only)"
  • itemURL: Direct link to the HathiTrust page-turner viewer
  • enumcron: Volume/issue info (e.g., "vol. 3") or false

Python Script

Use scripts/hathitrust_api.py for programmatic access (zero dependencies):

from scripts.hathitrust_api import HathiTrustAPI
ht = HathiTrustAPI()

# Single lookup
record = ht.lookup_isbn("0140449264")

# Multiple identifiers at once
results = ht.batch_lookup([
    ("isbn", "0140449264"),
    ("oclc", "228668653"),
    ("lccn", "96017900"),
])

# Get full MARC-XML
record = ht.lookup_isbn("0140449264", full=True)
marc = ht.get_marc_xml(record)

# Check if digitized and accessible
for item in ht.get_items(record):
    print(f"{item['orig']}: {item['usRightsString']} — {item['itemURL']}")

# Summarize
print(ht.summarize(record))

Typical Workflow: Cross-Reference with Other Catalogs

# 1. Search Harvard for a title
from harvard_api import HarvardLibraryAPI
harvard = HarvardLibraryAPI()
results = harvard.search(title="dream of the red chamber")
isbn = harvard.get_identifiers(results[0]).get("isbn")

# 2. Check HathiTrust for digital version
from hathitrust_api import HathiTrustAPI
ht = HathiTrustAPI()
record = ht.lookup_isbn(isbn)
if record and ht.has_full_view(record):
    print("Full text available:", ht.get_items(record)[0]["itemURL"])

API Etiquette

  • No authentication required
  • No documented rate limit, but the API is meant for small-batch lookups (not bulk harvesting)
  • Use HathiFiles or OAI-PMH for bulk data
  • CORS is fully enabled (access-control-allow-origin: *)

Related Skills

  • harvard-library-catalog: Search Harvard's catalog to find ISBNs/identifiers, then check HathiTrust for digital versions
  • wikidata-search: Cross-reference Wikidata identifiers (P243 = OCLC) with HathiTrust holdings

Resources

  • references/api_reference.md — Complete API spec with all identifier types and response fields
  • scripts/hathitrust_api.py — Python client with batch lookups, MARC extraction, and access checking