Vintage Typography Scans

1,200 high-res scans of vintage type specimens and letterpress ephemera, 1890s–1970s.

A scanned archive of historical typography: metal-type specimen books, letterpress posters, wood-type broadsides and packaging ephemera spanning eight decades. High-resolution and de-skewed, with era, technique and dominant-style metadata — a reference set for type design, style training and design history.

Dataset details

Supported tasks

• Typography and lettering style training
• Font-identification and historical-style classification
• Design reference and moodboarding

Dataset structure

Each item ships as a de-skewed scan plus catalogue metadata:

{
  "id": "type_00846",
  "image": "type_00846.tif",
  "era": "1930s",
  "technique": "wood_type_letterpress",
  "style": "condensed_sans",
  "source": "theatre_broadside",
  "dominant_colors": ["#1a1a1a", "#c0392b"]
}

Samples

Curation

Physical pieces from the creator's collection and partner archives were scanned at 600 DPI, de-skewed and colour-profiled, then catalogued by era and printing technique. Only public-domain or owned material is included.

Personal and sensitive information

None — printed matter only. No personal data; any names appearing are historical print content.

Considerations & limitations

• Paper aging (foxing, yellowing) is preserved, not corrected — factor it into training.
• Coverage skews Latin-script Western typography.
• Some ephemera lack firm dating and carry an estimated era.

License & provenance

Scanned from public-domain and owned print material. The sample archive (full_dataset.zip) is a free 50-item preview; the full set of 1,200 items unlocks with a purchase. Redistribution of the raw dataset is not permitted under either license.

Citation

@misc{nova-labs_00000015,
  title  = {Vintage Typography Scans},
  author = {Nova Labs (@nova-labs)},
  year   = {2026},
  note   = {TensorForge marketplace dataset, LST-00000015-0000-4000-8000-000000000015}
}