Next free-source hunt
We check for new free data every weekday.
Datasets / OSINT
static.case.law/

Harvard Caselaw Access Project

Freeauth: noneCC0-1.0#osint#historical-data#alt-data✓ verified Nov 1, 2026

Harvard Caselaw Access Project

Static bulk files of official, book-published United States state, federal and territorial court decisions, with JSON case text and metadata, HTML, per-volume ZIP files and scanned-volume PDFs. Case metadata includes party names, docket number, citation, court and decision date; the files download without a key.

Required Attribution:

Requested: "Caselaw Access Project. (2024). Retrieved [date], from [url]."

Coverage

The collection runs from 1658 through volumes published in 2020 and covers United States state, federal and listed territorial courts. It represents roughly 40 million pages in about 40,000 bound volumes; the site is static and the provider says the dataset is not kept current.

Verdict

A deep historical legal signal for tracking doctrine, litigation language, parties and court decisions across United States jurisdictions. It is static through 2020, excludes many unpublished and digital-only decisions, and is unsuitable for current legal status. Metadata received review, but case text remains raw OCR and can contain extraction errors. Store reporter, volume and citation identifiers, and preserve the requested project attribution even though CC0 imposes no binding attribution condition.

Get the free weekly email

Free weekly email of newly verified free data. Unsubscribe by emailing us.

Use in your agent

Load CAP reporter metadata into pandas and Parquet. Checked against the provider's documentation on Nov 1, 2026.

Raw skill file (/skills/caselaw-access-project.md)Also in our MCP server: call get_ingestion_skill with {"slug": "caselaw-access-project"}.

Access

  • No account or key. Static JSON metadata, volume ZIPs, case HTML and PDFs are public downloads.
  • This example uses the small root reporter-metadata file; full case text is partitioned by reporter and volume.

Request

bash
curl -O 'https://static.case.law/ReportersMetadata.json'

Format

JSON list of reporters with IDs, names, year range, jurisdiction objects, Harvard catalog IDs and reporter slug.

Load into pandas and Parquet

Needs Python 3 with requests, pandas and pyarrow.

python
"""CAP reporter metadata -> pandas -> Parquet. No key."""
import json
import requests
import pandas as pd

URL = "https://static.case.law/ReportersMetadata.json"
r = requests.get(URL, timeout=60)
r.raise_for_status()
df = pd.DataFrame(r.json())
df["jurisdictions"] = df["jurisdictions"].map(json.dumps)
df["harvard_hollis_id"] = df["harvard_hollis_id"].map(json.dumps)
df.to_parquet("cap_reporters_metadata.parquet", index=False)
print(len(df), "reporters", int(df["start_year"].min()), "to",
      int(df["end_year"].max()))

Run by us on 2026-11-01 (python 3.13.5 | pandas 3.1.0 | pyarrow 26.0.0 | requests 2.34.2). Result: 401 reporters 1658 to 2020

Limits

  • Rate limit: Not stated.
  • Coverage ends with volumes published in 2020 and is not kept current.
  • Full bulk volume files can be large; the checked root VolumesMetadata.json was about 30 MB.

Quirks

  • Reporter and jurisdiction fields are nested JSON; serialize or normalize them before writing a flat Parquet table.
  • Case text is raw OCR and can contain recognition errors even where key metadata was reviewed.
  • PDFs with selectable OCR are offered through 2018, while structured case data extends through 2020.
  • The collection omits many unpublished, trial-document and digital-only decisions.

Terms and attribution

  • Caselaw data and metadata are marked CC0 1.0 Universal and provided as is.
  • Harvard requests attribution and freely available improvements as nonbinding community norms.

Sources

Information and education only. Not financial, investment, legal or tax advice. Listings can be inaccurate, incomplete or out of date and are provided "as is". Free tiers, limits and licenses belong to each provider and can change, so check the provider's own terms before you rely on them. Terms & Disclaimer