This dataset is an Early-Access release shared by the Institutional Data Initiative for research and public-interest use (the “Service”). These terms are intended to support experimentation while encouraging collaboration and feedback as we refine the dataset and work with contributing institutions to define shared, long-term norms for open data reuse. To share questions or feedback, contact us at contact@institutional.org.

By accessing or downloading the dataset or otherwise using the Service, you agree to the following:

Noncommercial Use Only
You may use the Service solely for noncommercial purposes. Open-source projects and other public-use efforts are welcome, even if they may indirectly support commercial use, so long as they are unaffiliated with commercial actors or intent.

If you are affiliated with a commercial organization or plan to use the Service for commercial purposes (including AI model training), you will contact us first at contact@institutional.org.
No Redistribution
You may not share or redistribute the Service or any of the data provided through the Service, in whole or in part, including through public repositories or aggregators. If you want others to access it, please direct them to the attribution link.
Derivative Works
You may create derivative works for noncommercial use, but you may not make available any such derivative works that substantially reproduce the original dataset. Only outputs that are significantly transformed and cannot substitute for the original—such as evaluations, summary statistics, or visualizations—may be shared, with attribution.
Attribution
If you use the dataset in public-facing work, you must include attribution substantially similar to:

Institutional Newspapers provided by the Institutional Data Initiative with source material contributed by the Boston Public Library, available at https://institutional.org.

Minor modifications to fit citation style or formatting are permitted, provided the essential elements remain intact.
Provisional Terms
These terms apply only to this Early-Access release and may change. We are actively working with contributing institutions to develop a long-term framework for responsible, open data sharing.
DISCLAIMER OF WARRANTY
TO THE FULLEST EXTENT PERMITTED BY APPLICABLE LAW, ACCESS TO “SERVICE IS PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND (EXPRESS, IMPLIED, OR OTHERWISE), INCLUDING, WITHOUT LIMITATION, ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NONINFRINGEMENT. WE DO NOT WARRANT THAT THE SERVICE WILL OPERATE IN AN UNINTERRUPTED OR ERROR-FREE MANNER OR THAT THE SERVICE IS FREE OF VIRUSES OR OTHER HARMFUL COMPONENTS. WITHOUT LIMITING THE FOREGOING, WE DO NOT WARRANT THAT (A) THE SERVICE WILL MEET YOUR REQUIREMENTS OR EXPECTATIONS OR ACHIEVE THE INTENDED PURPOSES; (B) THE SERVICE WILL NOT EXPERIENCE OUTAGES OR OTHERWISE WILL BE UNINTERRUPTED, TIMELY, OR SECURE; (C) THE INFORMATION OR SERVICES OBTAINED THROUGH OR FROM THE SERVICE WILL BE ACCURATE, COMPLETE, CURRENT, ERROR-FREE, OR RELIABLE; (D) ANY DEFECTS IN OR ON THE SERVICE WILL BE CORRECTED; OR (E) THAT ANY POLLS OR OTHER SOLICITATIONS OF INFORMATION POSTED THROUGH THE SERVICE BY YOU OR OTHER USERS ARE SAFE OR APPROPRIATE FOR YOUR OR OTHER USERS’ PARTICIPATION. WE MAKE NO REPRESENTATION OR WARRANTY REGARDING YOUR ABILITY TO TRANSMIT AND RECEIVE INFORMATION FROM OR THROUGH THE SERVICE, AND YOU AGREE AND ACKNOWLEDGE THAT YOUR ABILITY TO ACCESS THE SERVICE MAY BE IMPAIRED.
LIMITATION OF LIABILITY
EXCEPT INSOFAR AS THE FOLLOWING LIMITATION MAY BE PROHIBITED BY APPLICABLE LAW, WE SHALL NOT BE LIABLE TO YOU OR TO ANY THIRD PARTY FOR ANY DIRECT, CONSEQUENTIAL, INDIRECT, PUNITIVE, SPECIAL, OR INCIDENTAL DAMAGES, WHETHER FORESEEABLE OR UNFORESEEABLE (INCLUDING, BUT NOT LIMITED TO, LOSS OF PROFITS OR EARNING POWER, LOSS OF DATA, LOSSES DUE TO ERRORS OR INTERRUPTION IN AVAILABILITY OF THE SERVICE, UNAVAILABILITY OF ANY SERVICE, SERVER, OR COMMUNICATIONS FACILITY, OR DAMAGES DUE TO ACTS OR OMISSIONS OF OTHERS USING THE SERVICE), ARISING OUT OF OR RELATING TO THE SERVICE, INCLUDING WITHOUT LIMITATION YOUR AND OTHERS’ USE OF OR INABILITY TO USE THE SERVICE, OR YOUR RELIANCE UPON INFORMATION OBTAINED FROM OR THROUGH THE SERVICE, WHETHER BASED IN CONTRACT, TORT, STATUTORY, OR OTHER LAW. OUR TOTAL CUMULATIVE LIABILITY TO YOU ARISING OUT OF OR RELATED TO THE SERVICE (INCLUDING, WITHOUT LIMITATION, IN THE WAYS DESCRIBED IN THE PRECEDING SENTENCE), WHETHER BASED IN CONTRACT, TORT, STATUTORY, OR OTHER LAW, WILL NOT EXCEED THE AMOUNT, IF ANY, THAT YOU PAID US TO USE THE SERVICE IN THE TWELVE MONTHS PRECEDING THE CLAIM. THE DISCLAIMERS AND LIMITATIONS SET FORTH IN THIS SECTION SHALL APPLY, TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, WHETHER OR NOT WE HAVE BEEN NEGLIGENT OR OTHERWISE AT FAULT.
YOU ACKNOWLEDGE THAT, FOR PURPOSES OF THE FOREGOING DISCLAIMERS AND LIMITATIONS, AS WELL AS THE INDEMNITY PROVISION IN SECTION 8 BELOW, THE TERMS “WE,” “OUR,” “US,” “INSTITUTIONAL DATA INITIATIVE,” AND “IDI” INCLUDE THE CORPORATE BODY PRESIDENT AND FELLOWS OF HARVARD COLLEGE, ALSO KNOWN AS HARVARD UNIVERSITY, AND ITS VARIOUS SCHOOLS, THE MEMBERS OF ITS GOVERNING BOARDS, AND ITS OFFICERS, FACULTY MEMBERS, EMPLOYEES, FELLOWS, AND TO THE EXTENT WORKING ON IDI, ITS STUDENTS, CONTRACTORS, AND REPRESENTATIVES.
Indemnification
You agree to indemnify us and hold us harmless from any and all claims, liabilities, damages, losses and expenses, including reasonable attorneys’ fees and costs, relating to or arising out of (a) your use or attempted use of the Service in violation of these Terms of Service; or (b) your violation of any law or rights of any third party in connection with your use of the Service.
Governing Law/ Jurisdiction
You agree that the Terms of Service and any claim or dispute arising out of or relating to the Service or Terms of Service will be governed by the laws of the Commonwealth of Massachusetts, excluding its conflicts of laws principles. You agree that all such claims and disputes will be heard and resolved exclusively in the federal or state courts located in and serving Middlesex or Suffolk County, Massachusetts, U.S.A. You consent to the personal jurisdiction of those courts over you for this purpose, and you waive and agree not to assert any objection to such proceedings in those courts (including any defense or objection of lack of proper jurisdiction or inconvenience of forum).
Whole Agreement/ Amendment
These Terms of Service constitute the entire agreement between you and Harvard with respect to your use of the Service. We reserve the right to amend these Terms of Service at any time. The Service will post notice of changes to the terms on this webpage, and by accessing the Service after modifications to these Terms of Service have been posted, you agree to be bound by all the modified terms. Accordingly, you should periodically revisit this page to review the then-current Terms of Service.

📰 Institutional Newspapers: Boston Public Library

A structured dataset derived from the Boston Public Library's public domain newspapers collection, produced by the Institutional Data Initiative in collaboration with Boston Public Library.

1,473,635 public domain newspaper scans, published between 1795 and 1930
83,147,041 individual crops segmented from those scans
16.3 billion o200k_base tokens of VLM OCR text, and 14.7 billion from Tesseract
Data for each crop: bbox coordinates, OCR, text analysis, crop type classification, language detection, NER, subject classification, reading order detection and text + image embeddings.

The Institutional Data Initiative at the Harvard Law School Library works with knowledge institutions—from libraries and museums to cultural groups and government agencies—to refine and publish their collections as data. Reach out to collaborate on your collections.

More information:

📄 Full analysis available in our technical report

See also:

Quickstart

One row is one newspaper scan. The dataset ships as a single train split.

from datasets import load_dataset

ds = load_dataset(
    "institutional/institutional-newspapers-bpl",
    split="train",
    streaming=True,
    token=True,
)

row = next(iter(ds))

print(row["issue_id_src"], row["page_number_gen"], row["year_ext"])
print(row["scan_image"])  # a PIL image, decoded from the embedded WEBP bytes

# Every crop_* column is a list with one entry per crop. They share a single
# index, and they are already sorted by the reading order detected for the page.
for bbox, category, text in zip(
    row["crop_bbox_gen"],
    row["crop_classification_gen"],
    row["crop_vlm_ocr_gen"],
):
    # One value per crop, NULL where the step produced no result.
    print(category or "", bbox, text or "")

Filtering preserves reading order, because it only drops entries from lists that are already sorted, so the surviving crops stay in the order in which the page reads.

# Example: Filtering out everything besides "Content" and "Section heading" crops while preserving reading order
KEEP = {"Content", "Section heading"}

article = "\n\n".join(
    text
    for category, text in zip(row["crop_classification_gen"], row["crop_vlm_ocr_gen"])
    if category and category[0] in KEEP and text
)

↑ Back to top

Loading only specific columns

Metadata can be read from the Parquet shards directly with column projection. Parquet is columnar, so the bytes of the skipped columns are never transferred.

This technique can help skip heavy columns when they are not needed (e.g, images, embeddings ...).

import pyarrow.parquet as pq
from huggingface_hub import HfFileSystem

SKIP = {
    "scan_image",
    "crop_tesseract_ocr_gen",
    "crop_text_embeddings",
    "crop_image_embeddings",
}

fs = HfFileSystem(token=True)
shards = fs.glob("datasets/institutional/institutional-newspapers-bpl/**/*.parquet")

with fs.open(shards[0], cache_type="none") as handle:
    parquet_file = pq.ParquetFile(handle, pre_buffer=True)
    keep = [
        field.name
        for field in parquet_file.schema_arrow
        if field.name not in SKIP
    ]
    table = parquet_file.read(columns=keep)

print(table.num_rows, table.column_names)

Shards are named BPL-part-NNNNN.parquet and hold up to 250 scans each. They are compressed with zstd and written with a row-group size of 10.

↑ Back to top

Fields

Suffixes:

Suffix	Description
`_src`	"From source". This field's data comes from information we gathered from the collection itself
`_ext`	"External". This field's data was pulled from an external source via a records matching mechanism.
`_gen`	"Generated". This field's data was generated as part of our analysis / processing pipeline.
`_exp`	"Experimental". Similar to `_gen`, but was generated through more experimental or exploratory means.

Setup:

One value per row. Each row is one scan (one page of one issue).
crop_* fields contain a list with one entry per crop. All of these lists share a single index, so index i refers to the same crop in every one of them, and they are all sorted by the reading order that the pipeline detected for the page. zip them to reassemble a crop, as in the Quickstart.

Field	Type	Role	Technical Report Section
scan_image	bytes	The source scan, stored as embedded WEBP bytes at quality 70.	1
corpus	string	Corpus identifier	13
issue_id_src	string	Issue identifier	2
newspaper_id_src	string	Newspaper (title) identifier	2
newspaper_id_type_gen	string	Type of the newspaper identifier.	2
page_number_gen	int	Page number within the issue. Inferred from filenames within source archive	2
scan_filename_src	string	Source scan filename	2
scan_width_src	int	Scan width in pixels	3
scan_height_src	int	Scan height in pixels	3
year_ext	int	Publication year	2
month_ext	int	Publication month	2
day_ext	int	Publication day	2
edition_ext	int	Edition	2
metadata_source_gen	string	Source of the issue-level metadata	5, 13
city_gen	string	Locality city (post-processed, sourced from metadata_source_gen)	13
state_gen	string	Locality state (post-processed, sourced from metadata_source_gen)	13
country_gen	string	Locality country (post-processed, sourced from metadata_source_gen)	13
language_ext	string	Issue-level language code (sourced from metadata_source_gen)	5
crop_bbox_gen	list[list[float]]	Crop bounding boxes	3
crop_bbox_conf_gen	list[float]	Crop detection confidence scores	3
crop_vlm_ocr_gen	list[string]	dots.mocr OCR text (post-processed)	4
crop_tesseract_ocr_gen	list[struct]	OCR text of each crop from Tesseract 5, as {text, metadata}. text is plain text. metadata is a list holding one record per recognized word, each with text (the surface form), conf (the Tesseract confidence score), and bbox_xyxy (the word box as [x_min, y_min, x_max, y_max], in pixel coordinates relative to the crop, rescaled back to the crop's full resolution). Useful for building OCR outputs compatible with existing library discovery systems.	4
crop_vlm_ocr_token_count_gen	list[int]	dots.mocr o200k_base token count	4
crop_tesseract_ocr_token_count_gen	list[int]	Tesseract o200k_base token count	4
crop_text_analysis_gen	list[list[string]]	Per-crop text metrics. Holds tokenizability_score, char_count, word_count, word_count_unique, word_type_token_ratio, sentence_count, and sentence_count_unique, each prefixed tesseract_ and vlm_, plus vlm_has_table and vlm_has_markdown.	5
crop_classification_gen	list[string]	Final crop-type category	6
crop_classification_image_only_gen	list[string]	Image-classifier category	6
crop_classification_image_only_conf_gen	list[float]	Image-classifier confidence	6
crop_classification_text_only_gen	list[string]	Text-classifier category	6
crop_classification_text_only_conf_gen	list[float]	Text-classifier confidence	6
crop_language_gen	list[string]	Detected language (post-processed, issue-level language used if confidence is low)	5
crop_language_conf_gen	list[float]	Detected language confidence (post-processed, NULL if issue-level language was used)	5
crop_ner_per_gen	list[list[string]]	Person entities	8
crop_ner_per_conf_gen	list[list[float]]	Person entity confidence	8
crop_ner_loc_gen	list[list[string]]	Location entities	8
crop_ner_loc_conf_gen	list[list[float]]	Location entity confidence	8
crop_ner_org_gen	list[list[string]]	Organization entities	8
crop_ner_org_conf_gen	list[list[float]]	Organization entity confidence	8
crop_subject_gen	list[list[string]]	Subject ranking	9
crop_subject_conf_gen	list[list[float]]	Subject confidence	9
crop_chronam_thesauri_matches_exp	list[struct]	Keyword matches from the Race, Ethnicity, Citizenship and Immigration Keyword Thesauri for Chronicling America. Holds matches (a list of {category, terms}, terms being {term, count} pairs), match_count, and term_count, each prefixed tesseract_ and vlm_. Naive match: may be used for downstream research, but not "as is". (See section 11)	11
crop_text_embeddings	list[list[float]]	Text embeddings (potion-multilingual, 256-d)	10
crop_image_embeddings	list[list[float]]	Image embeddings (DINOv2-small, 384-d)	1

↑ Back to top

Terms of Use

This dataset is shared by the Institutional Data Initiative for research and public-interest use (the “Service”). These terms are intended to support experimentation while encouraging collaboration and feedback as we refine the dataset and work with contributing institutions to define shared, long-term norms for open data reuse. To share questions or feedback, contact us at contact@institutional.org.

By accessing or downloading the dataset or otherwise using the Service, you agree to the following:

Noncommercial Use Only
You may use the Service solely for noncommercial purposes. Open-source projects and other public-use efforts are welcome, even if they may indirectly support commercial use, so long as they are unaffiliated with commercial actors or intent.

If you are affiliated with a commercial organization or plan to use the Service for commercial purposes (including AI model training), you will contact us first at contact@institutional.org.
No Redistribution
You may not share or redistribute the Service or any of the data provided through the Service, in whole or in part, including through public repositories or aggregators. If you want others to access it, please direct them to the attribution link.
Derivative Works
You may create derivative works for noncommercial use, but you may not make available any such derivative works that substantially reproduce the original dataset. Only outputs that are significantly transformed and cannot substitute for the original—such as evaluations, summary statistics, or visualizations—may be shared, with attribution.
Attribution
If you use the dataset in public-facing work, you must include attribution substantially similar to:

Institutional Newspapers provided by the Institutional Data Initiative with source material contributed by Boston Public Library, available at https://institutional.org.

Minor modifications to fit citation style or formatting are permitted, provided the essential elements remain intact.
Provisional Terms
These terms apply only to this release and may change. We are actively working with contributing institutions to develop a long-term framework for responsible, open data sharing.
DISCLAIMER OF WARRANTY
TO THE FULLEST EXTENT PERMITTED BY APPLICABLE LAW, ACCESS TO “SERVICE IS PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND (EXPRESS, IMPLIED, OR OTHERWISE), INCLUDING, WITHOUT LIMITATION, ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NONINFRINGEMENT. WE DO NOT WARRANT THAT THE SERVICE WILL OPERATE IN AN UNINTERRUPTED OR ERROR-FREE MANNER OR THAT THE SERVICE IS FREE OF VIRUSES OR OTHER HARMFUL COMPONENTS. WITHOUT LIMITING THE FOREGOING, WE DO NOT WARRANT THAT (A) THE SERVICE WILL MEET YOUR REQUIREMENTS OR EXPECTATIONS OR ACHIEVE THE INTENDED PURPOSES; (B) THE SERVICE WILL NOT EXPERIENCE OUTAGES OR OTHERWISE WILL BE UNINTERRUPTED, TIMELY, OR SECURE; (C) THE INFORMATION OR SERVICES OBTAINED THROUGH OR FROM THE SERVICE WILL BE ACCURATE, COMPLETE, CURRENT, ERROR-FREE, OR RELIABLE; (D) ANY DEFECTS IN OR ON THE SERVICE WILL BE CORRECTED; OR (E) THAT ANY POLLS OR OTHER SOLICITATIONS OF INFORMATION POSTED THROUGH THE SERVICE BY YOU OR OTHER USERS ARE SAFE OR APPROPRIATE FOR YOUR OR OTHER USERS’ PARTICIPATION. WE MAKE NO REPRESENTATION OR WARRANTY REGARDING YOUR ABILITY TO TRANSMIT AND RECEIVE INFORMATION FROM OR THROUGH THE SERVICE, AND YOU AGREE AND ACKNOWLEDGE THAT YOUR ABILITY TO ACCESS THE SERVICE MAY BE IMPAIRED.
LIMITATION OF LIABILITY
EXCEPT INSOFAR AS THE FOLLOWING LIMITATION MAY BE PROHIBITED BY APPLICABLE LAW, WE SHALL NOT BE LIABLE TO YOU OR TO ANY THIRD PARTY FOR ANY DIRECT, CONSEQUENTIAL, INDIRECT, PUNITIVE, SPECIAL, OR INCIDENTAL DAMAGES, WHETHER FORESEEABLE OR UNFORESEEABLE (INCLUDING, BUT NOT LIMITED TO, LOSS OF PROFITS OR EARNING POWER, LOSS OF DATA, LOSSES DUE TO ERRORS OR INTERRUPTION IN AVAILABILITY OF THE SERVICE, UNAVAILABILITY OF ANY SERVICE, SERVER, OR COMMUNICATIONS FACILITY, OR DAMAGES DUE TO ACTS OR OMISSIONS OF OTHERS USING THE SERVICE), ARISING OUT OF OR RELATING TO THE SERVICE, INCLUDING WITHOUT LIMITATION YOUR AND OTHERS’ USE OF OR INABILITY TO USE THE SERVICE, OR YOUR RELIANCE UPON INFORMATION OBTAINED FROM OR THROUGH THE SERVICE, WHETHER BASED IN CONTRACT, TORT, STATUTORY, OR OTHER LAW. OUR TOTAL CUMULATIVE LIABILITY TO YOU ARISING OUT OF OR RELATED TO THE SERVICE (INCLUDING, WITHOUT LIMITATION, IN THE WAYS DESCRIBED IN THE PRECEDING SENTENCE), WHETHER BASED IN CONTRACT, TORT, STATUTORY, OR OTHER LAW, WILL NOT EXCEED THE AMOUNT, IF ANY, THAT YOU PAID US TO USE THE SERVICE IN THE TWELVE MONTHS PRECEDING THE CLAIM. THE DISCLAIMERS AND LIMITATIONS SET FORTH IN THIS SECTION SHALL APPLY, TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, WHETHER OR NOT WE HAVE BEEN NEGLIGENT OR OTHERWISE AT FAULT.
YOU ACKNOWLEDGE THAT, FOR PURPOSES OF THE FOREGOING DISCLAIMERS AND LIMITATIONS, AS WELL AS THE INDEMNITY PROVISION IN SECTION 8 BELOW, THE TERMS “WE,” “OUR,” “US,” “INSTITUTIONAL DATA INITIATIVE,” AND “IDI” INCLUDE THE CORPORATE BODY PRESIDENT AND FELLOWS OF HARVARD COLLEGE, ALSO KNOWN AS HARVARD UNIVERSITY, AND ITS VARIOUS SCHOOLS, THE MEMBERS OF ITS GOVERNING BOARDS, AND ITS OFFICERS, FACULTY MEMBERS, EMPLOYEES, FELLOWS, AND TO THE EXTENT WORKING ON IDI, ITS STUDENTS, CONTRACTORS, AND REPRESENTATIVES.
Indemnification
You agree to indemnify us and hold us harmless from any and all claims, liabilities, damages, losses and expenses, including reasonable attorneys’ fees and costs, relating to or arising out of (a) your use or attempted use of the Service in violation of these Terms of Service; or (b) your violation of any law or rights of any third party in connection with your use of the Service.
Governing Law/ Jurisdiction
You agree that the Terms of Service and any claim or dispute arising out of or relating to the Service or Terms of Service will be governed by the laws of the Commonwealth of Massachusetts, excluding its conflicts of laws principles. You agree that all such claims and disputes will be heard and resolved exclusively in the federal or state courts located in and serving Middlesex or Suffolk County, Massachusetts, U.S.A. You consent to the personal jurisdiction of those courts over you for this purpose, and you waive and agree not to assert any objection to such proceedings in those courts (including any defense or objection of lack of proper jurisdiction or inconvenience of forum).
Whole Agreement/ Amendment
These Terms of Service constitute the entire agreement between you and Harvard with respect to your use of the Service. We reserve the right to amend these Terms of Service at any time. The Service will post notice of changes to the terms on this webpage, and by accessing the Service after modifications to these Terms of Service have been posted, you agree to be bound by all the modified terms. Accordingly, you should periodically revisit this page to review the then-current Terms of Service.

↑ Back to top

Disclaimers

Harmful Language and Content in this Dataset

This dataset is a collection of historical works that reflect the language, culture, and perspectives of their time. Users should be aware that some materials may contain language or portrayals that are outdated, offensive, or harmful today, such as racism, sexism, colonial attitudes, and other forms of discrimination. Some content may include inaccurate information, providing insight into historical contexts that existed at the time of writing. The materials are maintained in their original form to retain contextual understanding and facilitate research efforts, but we encourage critical awareness and cultural sensitivity for the creators and/or subjects of the collection. These materials are offered as part of a historical perspective, but should not be considered a stand-alone research collection constructed to give a balanced perspective on any topic.

Harmful Language in Bibliographic Description

Metadata for this collection may contain language that is overtly or implicitly harmful, outdated, or biased, or may by omission fail to represent important perspectives. Metadata may contain language created decades ago. It is common practice within the field of library science to reuse descriptions provided from the creator of the materials. While in some instances this allows communities and individuals to represent their materials in their own words, unexamined use of this practice may mean that racist or other offensive terminologies appear in our description. We also use national standardized terms in our work that can be outdated and harmful. Note that terminology in historical materials and in library descriptions does not always match the language we currently understand to be preferred by members of the communities depicted.

Furthermore, we acknowledge that the act of collecting materials is not always neutral, and the work of describing and classifying library materials is influenced by inherent personal, institutional, and societal biases. Outdated or offensive terminologies may be present in metadata such as subject headings, and harmful language or bias may be introduced by catalogers supplying titles and descriptions. In other cases, the source materials themselves present racist, offensive or otherwise harmful viewpoints in titles or descriptions that are routinely transcribed by catalogers.

Note: Some language in this statement was adopted from Harvard Library's statement on Harmful Language in Library collections.

Generated and Experimental Content

This dataset contains generated and/or experimental content. While reasonable care was taken to ensure its quality, it is provided "as is," without warranties of any kind. It may contain errors or inaccuracies; users should verify the data independently and apply their own judgment.

The fields carrying the _exp suffix are experimental by definition. We also treat the named-entity and subject fields as experimental, as noted in Crop-level list fields.

Rights Determination

We respect the intellectual property rights of authors, publishers, and other rights holders. We have taken deliberate steps to include only those issues for which there is no known copyright restriction. The source materials used in the preparation of this dataset were published before 1931, making them public domain in the United States. Issue-level metadata provided by Boston Public Library was used to make this assessment.

While this is relatively low risk, some materials in this dataset may be in the public domain in the United States but still subject to copyright or other rights protections in other jurisdictions. Additionally, the absence of an explicit copyright claim or rights status does not guarantee that a work is in the public domain, either in the U.S. or abroad. Information about the copyright status of individual issues is provided on a good-faith basis and reflects available data at the time of determination, but we cannot guarantee its completeness or accuracy.

Users of this dataset will be solely responsible for making independent legal assessments about how and where they use the materials. Some uses of materials may also be restricted by trademark, privacy, publicity rights, or other such rights or restrictions. It is the user's sole responsibility to consider the possibility that such rights or restrictions may be involved and to secure any needed permissions. If any rights holder believes that a work included in this release is misidentified or improperly included, we welcome contact and will promptly review any concerns. Our goal is to provide broad public access while maintaining respect for intellectual property rights and ensuring responsible data stewardship.

↑ Back to top

Cite

@misc{cargnelutti2026institutionalnewspaperspipelinederiving,
      title={Institutional Newspapers Pipeline: Deriving billions of high quality tokens from historical newspapers}, 
      author={Matteo Cargnelutti and Catherine Brobston and Eben English and Jake Sadow and Kacie Bailey and Greg Leppert and Amanda Watson and Jessica Chapel and Jonathan Zittrain},
      year={2026},
      eprint={2608.18972},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2608.18972}, 
}

Downloads last month: 85

Total file size:

9.93 TB

Models trained or fine-tuned on institutional/institutional-newspapers-bpl

Collection including institutional/institutional-newspapers-bpl

Institutional Newspapers

Collection

A growing corpus of newspapers, parsed and optimized for computational access. • 6 items • Updated about 20 hours ago

Paper for institutional/institutional-newspapers-bpl

Institutional Newspapers Pipeline: Deriving billions of high quality tokens from historical newspapers

Paper • 2608.18972 • Published 3 days ago