Document type · reference
Reference Word documents
32,559 documents across 73 languages and 9 topics. See /classification for what this label covers.
32,559documents
73languages
9topics
What's in this slice
Product catalogs, directories, glossaries, FAQs, indexes, and other reference or lookup material.
Useful for: catalog-data extraction, glossary-building, FAQ retrieval, structured-lookup benchmarks.
By topic
| Topic | Count |
|---|---|
| Education | 7,746 |
| General | 6,959 |
| Government | 4,421 |
| Healthcare | 4,223 |
| Finance | 2,908 |
| Legal / Judicial | 2,268 |
| Environment | 1,717 |
| Technology | 1,186 |
| Nonprofit | 1,131 |
Top languages
| Lang | Count | Share |
|---|---|---|
| en | 15,855 | 52.1% |
| fr | 2,129 | 7.0% |
| cs | 1,448 | 4.8% |
| unknown | 1,397 | 4.6% |
| ru | 1,332 | 4.4% |
| la | 1,288 | 4.2% |
| + 67 more | ||
Share is computed against the top 20 languages for this type (30,426 docs), matching what the API returns. A handful of documents fall outside the top 20 or have no detected language.
Sample documents · real rows from the corpus
| ID | Filename | Topic | Lang | Conf |
|---|---|---|---|---|
| 5eff8f3b3a63 | WebsitesforScienceCurrentEvents.docx | Education | en | 0.98 |
| 4a4281f5025c | 954780 | Education | en | 0.98 |
| 257860130524 | filedownload.ashx | Education | en | 0.98 |
| b05065dd192a | 7ac6128e-b8a0-44f7-8cbd-69dd9452fdfd | Healthcare | pl | 0.98 |
| c7dccd7b1978 | filedownload.ashx | Education | en | 0.98 |
ID column shows the first 12 characters of the SHA-256 content hash; the full hash is the stable reference. Real public-web filenames vary widely: descriptive, numeric, or URL-fragment shaped.
Access
# All reference documents
curl "https://api.docxcorp.us/manifest?type=reference" -o reference-manifest.txt
# High-confidence English subset
curl "https://api.docxcorp.us/manifest?type=reference&lang=en&min_confidence=0.8"See /download for full access patterns.