Search loc collections

When the user asks to search the Library of Congress — books, photos, maps, manuscripts, audio — by keyword, format, or date range, reach for the loc.gov search endpoint. One unauthenticated GET with fo=json.

search-loc-collections · v2 · updated 2026-04-16

Agents: This page is a SKILL.md-style capability guide. For JSON, call GET /api/skills/search-loc-collections. To drop this into a local Claude Code install, copy the frontmatter + body below into ~/.claude/skills/search-loc-collections/SKILL.md.

When to use this skill

When the user asks to search the Library of Congress catalog — books, photographs, maps, manuscripts, audio recordings — by keyword, subject, date range, or format. LoC holds millions of catalog records including foreign-language holdings: a search for "constitution" returns the 1923 Romanian constitution alongside Pennsylvania's. For retrieving a known item by its LoC identifier, use the /item/{id}/ follow-up described below — this skill covers the search that gets you there.

Your best first call

curl "https://www.loc.gov/search/?q=illuminated+manuscript&fo=json&c=5"

No auth. No key. The one mandatory parameter is fo=json — without it, every LoC endpoint returns an HTML page, not JSON. q is the search query; c limits results per page.

The results array contains item objects with these key fields:

id — full LoC URL for the item (e.g. http://lccn.loc.gov/30027597). Extract the numeric or alphanumeric segment after the last / for the /item/{id}/ endpoint.
title, date — item title and publication date
original_format — array like ["book"] or ["photo, print, drawing"]
subject — array of subject headings (e.g. ["romania", "constitutions"])
digitized — boolean; true means image resources exist in the full item record
item.call_number, item.created_published — catalog metadata nested under item

Narrow further with facet filters: fa=original-format:book (format), dates=1900/1950 (date range). Combine q with one or two facets for the most targeted results.

Fallbacks (when the best call isn't enough)

Known item identifier (from a search result or a loc.gov URL) → https://www.loc.gov/item/{id}/?fo=json returns the full record including digitized image URLs in resources. Extract the numeric ID from a search result's id field — it is a full URL like http://lccn.loc.gov/30027597, not a slug.
No fallback if loc.gov is down — this is the only free, programmatic access to the Library of Congress catalog.

Pitfalls

fo=json is mandatory on every request. Without it, you get an HTML page. This is the single most common mistake and it is not prominently documented.
The id field is a full URL, not a slug. You cannot pass "http://lccn.loc.gov/30027597" directly to /item/{id}/. Extract the numeric segment (e.g. 30027597 or 2014711606) after the last /.
search.hits and pagination.total differ. hits is the unfiltered count across all formats; total is the count after your facet filters are applied. Use pagination.total for page navigation.
Cloudflare bot protection is aggressive. Rapid-fire requests return 403 with a JavaScript challenge page. Add a recognizable User-Agent header and pace requests.

One-line summary for the user

I can search the Library of Congress catalog — books, photos, maps, manuscripts — by keyword, format, and date range, but every request must include fo=json or you get HTML instead of JSON.

APIs this skill uses

Library of Congress APIs · primary · verified

The Library of Congress provides APIs for accessing its digital collections including books, photos, videos, maps, manuscripts, and more.

Generated from

Library of Congress APIs tutorial Getting Started with Library of Congress

SKILL.md source (frontmatter + body)

---
name: search-loc-collections
description: When the user asks to search the Library of Congress — books, photos, maps, manuscripts, audio — by keyword, format, or date range, reach for the loc.gov search endpoint. One unauthenticated GET with fo=json.
---

## When to use this skill

When the user asks to search the Library of Congress catalog — books, photographs, maps, manuscripts, audio recordings — by keyword, subject, date range, or format. LoC holds millions of catalog records including foreign-language holdings: a search for "constitution" returns the 1923 Romanian constitution alongside Pennsylvania's. For retrieving a known item by its LoC identifier, use the `/item/{id}/` follow-up described below — this skill covers the search that gets you there.

## Your best first call

```bash
curl "https://www.loc.gov/search/?q=illuminated+manuscript&fo=json&c=5"
```

No auth. No key. The one mandatory parameter is `fo=json` — without it, every LoC endpoint returns an HTML page, not JSON. `q` is the search query; `c` limits results per page.

The `results` array contains item objects with these key fields:

- `id` — full LoC URL for the item (e.g. `http://lccn.loc.gov/30027597`). Extract the numeric or alphanumeric segment after the last `/` for the `/item/{id}/` endpoint.
- `title`, `date` — item title and publication date
- `original_format` — array like `["book"]` or `["photo, print, drawing"]`
- `subject` — array of subject headings (e.g. `["romania", "constitutions"]`)
- `digitized` — boolean; `true` means image resources exist in the full item record
- `item.call_number`, `item.created_published` — catalog metadata nested under `item`

Narrow further with facet filters: `fa=original-format:book` (format), `dates=1900/1950` (date range). Combine `q` with one or two facets for the most targeted results.

## Fallbacks (when the best call isn't enough)

- **Known item identifier (from a search result or a loc.gov URL)** → `https://www.loc.gov/item/{id}/?fo=json` returns the full record including digitized image URLs in `resources`. Extract the numeric ID from a search result's `id` field — it is a full URL like `http://lccn.loc.gov/30027597`, not a slug.
- **No fallback if loc.gov is down** — this is the only free, programmatic access to the Library of Congress catalog.

## Pitfalls

- **`fo=json` is mandatory on every request.** Without it, you get an HTML page. This is the single most common mistake and it is not prominently documented.
- **The `id` field is a full URL, not a slug.** You cannot pass `"http://lccn.loc.gov/30027597"` directly to `/item/{id}/`. Extract the numeric segment (e.g. `30027597` or `2014711606`) after the last `/`.
- **`search.hits` and `pagination.total` differ.** `hits` is the unfiltered count across all formats; `total` is the count after your facet filters are applied. Use `pagination.total` for page navigation.
- **Cloudflare bot protection is aggressive.** Rapid-fire requests return 403 with a JavaScript challenge page. Add a recognizable `User-Agent` header and pace requests.

## One-line summary for the user

I can search the Library of Congress catalog — books, photos, maps, manuscripts — by keyword, format, and date range, but every request must include `fo=json` or you get HTML instead of JSON.

« Back to all skills