sample-first data packs
Five packs an agent can read without a browser: 149,673 rows of sports, housing, and bounty data plus a Playwright scraper pack. Every dataset answers curl with JSON, no key, no signup. The MCP server goes further and queries the full released files.
MCP server. Point your MCP client at npx --allow-git=all -y github:jayjex/dataset-mcp (the flag is for npm 12+, npm 10 and 11 run the plain form). Tools: list_datasets, get_dataset_info, get_sample, query_dataset, and get_stats. query_dataset filters and paginates the full released files, 100 rows per call, with =, contains, gt, and lt operators. The server caches the files locally and verifies SHA-256 on every load. Full setup and a captured session: MCP server for datasets.
query_dataset("hud-fmr-2026", {
where: [{ column: "state", op: "=", value: "TX" }],
limit: 5
})
# 3,247 Texas rows matched, returns zip, area_name, fmr_0br..fmr_4br
Plain HTTP. Every dataset exposes a sample endpoint an agent can fetch directly. sample.json returns the schema and the first rows, sample.csv returns the same rows as CSV:
curl -s https://jayjex.github.io/data-vault/data/airbnb-six-cities/sample.json | jq '.columns, .records[0]'
The machine-readable index of all five lives in catalog.json, and the agent docs cover the endpoints, a Python example, and the MCP config for Claude Desktop. Deciding between grabbing the file and querying it? CSV vs API for data gives the decision table. Want more than the endpoints? The free tools for AI agents roundup adds the registry and docs links we check ourselves.
7,548 NFL games with scores, closing spreads, totals, moneylines, odds, rest days, QBs, coaches, and stadium weather.
Full pack: on Getly.
FY2026 fair market rents for 51,895 ZIP codes, 3,229 counties, and 52 states, five bedroom sizes per row.
Full pack: on Getly.
90,169 Airbnb listings across Austin, Nashville, Denver, New York City, Las Vegas, and San Diego in one normalized schema, host data stripped.
Full pack: on Getly.
28 open Superteam Earn listings with full description HTML, the raw cards dump, and a map of 186+ reverse-engineered API routes.
Full pack: on Getly. Query it free first: the Superteam Earn data guide shows real MCP output against the full files.
Ten self-contained Playwright scripts for pagination, infinite scroll, login sessions, price monitors, sitemap crawling, and retries.
Full pack: on Getly.
Samples are CC BY 4.0 unless the dataset page says otherwise. Republish them with credit: "Data Vault (jayjex.github.io/data-vault), CC BY 4.0", plus the upstream source noted per dataset (nflverse, HUD, Inside Airbnb, superteam.fun). Full packs carry their own license terms per purchase. The selling-process walkthrough shows where each credit line and license check happens.
Samples, aggregate CSVs, and every MCP query cost nothing and need no account. The full files ship as one release in the MCP repo, so a query against 90,169 rows downloads nothing on your side beyond the server's own cache. Buying a full pack gets you the raw CSVs for local work, useful when a pipeline needs the whole table rather than paged queries. The catalog page lists every pack with its row counts and niches.