About LCSH Tools
Making library cataloging accessible with artificial intelligence — validated against the world's most authoritative subject heading system.
What Are Library of Congress Subject Headings?
Library of Congress Subject Headings (LCSH) are a controlled vocabulary used by libraries worldwide to describe the subjects of books, journals, films, and other materials. Maintained by the Library of Congress since 1898, LCSH is the most widely adopted subject heading system in the world, used by libraries in over 130 countries.
When you search a library catalog for books about "climate change" or "artificial intelligence," you're relying on subject headings that catalogers carefully assigned to each item. These standardized terms ensure that materials on the same topic can be found together, regardless of the title or language of the work.
Assigning subject headings is a skilled, time-intensive process. Catalogers must understand the content of each work, identify its key themes, and select the most appropriate authorized headings from a vocabulary of over 340,000 terms. They must also follow complex rules about subdivisions, geographic qualifiers, and chronological terms.
The Vision
LCSH Tools was created to bridge the gap between artificial intelligence and library science standards. The goal is not to replace professional catalogers, but to augment their expertise — accelerating the most time-consuming parts of subject analysis while ensuring every suggestion is grounded in the Library of Congress authorities database.
The project follows a zero-knowledge architecture: all AI inference and Library of Congress lookups run directly in the user's browser. No bibliographic data or API keys ever leave the user's device (except to their chosen AI provider). This design puts catalogers in complete control of their data and their workflow.
By combining modern AI capabilities with real-time validation against official authorities, LCSH Tools produces results that are both innovative and standards-compliant — ready to be imported directly into any Integrated Library System.
The Ecosystem
LCSH Tools is a suite of three complementary applications, each designed for a different workflow:
Cataloging Assistant (PWA)
The flagship application. A full-featured Progressive Web App with a 3-step wizard, multi-provider AI support (OpenAI, Google Gemini, DeepSeek, Qwen, and more), dual authority validation against LCSH and LCNAF, MARC record generation, image analysis, and session history with CSV export. Installs on any device and works offline.
Learn more →Chrome Extension
A lightweight browser extension powered by Google Gemini. Generates LCSH suggestions, validates them against id.loc.gov by scraping search results in background tabs, calculates similarity scores, and produces MARC records — all from a browser sidebar without leaving your current workflow.
Learn more →MCP Server
A Model Context Protocol server that gives AI assistants like Claude direct access to the Library of Congress authorities database. Three specialized search tools (LCSH, LCSH keyword, LCNAF) let AI assistants autonomously look up and validate subject headings during conversation.
Learn more →Who Is This For?
Library Catalogers
Speed up subject heading assignment while maintaining accuracy. AI generates initial suggestions; you make the final call.
Library Science Students
Learn about LCSH, MARC records, and authority files through hands-on practice with real Library of Congress data.
Technical Services Librarians
Evaluate AI-assisted cataloging tools for your institution. Test multiple AI providers and customize cataloging rules.
Researchers & Archivists
Generate standardized subject headings for collections, research databases, and institutional repositories.
Research
The concepts behind LCSH Tools are described in the following paper:
Tang, K. L. & Jiang, Y. (2025). Better Recommendations: Validating AI-generated Subject Terms Through LOC Linked Data Service.
This paper investigates incorporating machine-generated subject terms into library cataloging systems with validation via the Library of Congress Linked Data Service. The authors propose a hybrid methodology merging AI capabilities with human review to improve metadata quality and cataloging efficiency.
Read on arXiv →Acknowledgments
LC Linked Data Service
All LCSH Tools are built on the LC Linked Data Service, which provides machine-readable access to the Library of Congress's authority files — including Subject Headings (LCSH) and Name Authorities (LCNAF). This service is the backbone of every validation, search, and authority lookup performed by the Cataloging Assistant, Chrome Extension, and MCP Server. We are grateful to the Library of Congress for making this data openly available through their public APIs.
Glossary
Key terms used throughout LCSH Tools.
| Term | Definition |
|---|---|
| LCSH | Library of Congress Subject Headings — a controlled vocabulary for describing library material subjects |
| LCNAF | Library of Congress Name Authority File — an authoritative list of name headings |
| MARC | MAchine-Readable Cataloging — a standard format for bibliographic information |
| 650 tag | MARC field for topical subject headings from LCSH |
| 600 tag | MARC field for personal name subject headings from LCNAF |
| 610 tag | MARC field for corporate name subject headings from LCNAF |
| PWA | Progressive Web App — a web app that can be installed like a native app |
| LOC | Library of Congress |
| ILS | Integrated Library System — software for managing library cataloging and circulation |
| MCP | Model Context Protocol — a standard for connecting AI assistants to external tools and data |
| Similarity Score | A percentage (0–100%) indicating match quality against LOC headings, calculated using Levenshtein distance |
| API Key | A secret token that authenticates requests to an AI provider |