claimscout dataset

A daily-refreshed registry of every active US class-action settlement and mass-tort matter, with eligibility hints, payout bands, claim deadlines, multi-source signal velocity, and provenance to primary court filings.

License: CC-BY-4.0. Free to use, redistribute, and remix — including commercially. Attribution to claimscout required. See full text at creativecommons.org/licenses/by/4.0.

Downloads

FileFormatDescription
active-settlements.csv CSV One row per active matter. Refreshed query-time from Postgres.
signals.csv CSV Last 90 days of underlying signals (news, court filings, regulatory).
/api/v1/matters JSON Live API. OpenAPI spec at /openapi.json.

Citation

claimscout. Active US class-action settlements registry.
Hugging Face Datasets, 2026.
https://huggingface.co/datasets/claimscout/active-settlements

Quick pulls

# pandas
import pandas as pd
df = pd.read_csv("https://getclaimscout.com/dataset/active-settlements.csv")

# curl
curl -sL https://getclaimscout.com/dataset/active-settlements.csv > settlements.csv

# duckdb
SELECT slug, name, status, payout_band, claims_close_iso
FROM read_csv_auto('https://getclaimscout.com/dataset/active-settlements.csv')
WHERE status = 'claims_open'
ORDER BY claims_close_iso ASC
LIMIT 20;

Refresh + reproducibility

The CSV streams live from PG (1-hour edge cache). For point-in-time snapshots, use the Hugging Face mirror at huggingface.co/datasets/claimscout/active-settlements, which captures a daily-tagged version. Each daily revision is preserved.

Schema

Full schema documented in the Hugging Face dataset card. Each column is documented; lists are JSON-encoded for CSV-safety.

Agent-ready

Building an LLM tool, an MCP server, or an automated lawyer-tech workflow? See /llms.txt for the agent quickstart, /openapi.json for the full contract, and pip install claimscout-mcp for the stdio MCP server.