
NYSgpt Data Library
com.nysgptPublished by nysgpt.comOfficial publisherv2.0.0Updated Nov 17, 2026
SQL over 52.8B public records: USAspending, OpenAlex, patents, LEHD jobs, SEC, EPA, FERC, Medicare.
Overview
Remote MCP server that lets an assistant run read-only SQL over 52.8 billion public records in 6,396 preloaded tables, from federal spending to patents and…
- What you can do
- Discover the available subject databases and their tables, with row counts
- Inspect a table's fields, types and description before querying
- Run one read-only SQL statement (Athena/Trino dialect) per call, returning up to 1,000 rows
- Query preloaded public datasets such as USAspending, OpenAlex, patents, LEHD jobs, SEC filings, EPA, FERC and Medicare
- When to use it
- Useful when you need to answer questions over large U.S. public datasets without downloading or restoring archives, such as federal spending, patent, scholarly, labor, energy, health or court records.
- Requirements
- Remote streamable HTTP endpoint; nothing to install. An access key is required, passed as a key query parameter or the X-Api-Key header. Access is paid, starting at a day pass, with daily query and data-scanned limits.
Installation
In SourceWeft
- Open NYSgpt Data Library in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.
Use with an agent
Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.
Read https://sourceweft.com/mcp/com.nysgpt%2Fdata-library/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.Other MCP clients
Add this to your client's mcpServers config.
{
"mcpServers": {
"nysgpt-data-library": {
"type": "http",
"url": "https://data.nysgpt.com/mcp"
}
}
}README
NYSgpt Data Library
52.8B public records in 6,396 SQL tables, queryable from Claude (or any MCP client) and over plain HTTP. Federal spending, patents, company filings, scholarly works, labor, census, energy, health, courts and more, already loaded and typed. Nothing to download or restore.
- Browse every dataset: https://nysgpt.com/data
- MCP server:
https://data.nysgpt.com/mcp(remote, streamable HTTP, no install; needs a key) - Access: by key, from $29: https://nysgpt.com/data#pricing. Free 20,000-row samples of ten datasets: https://huggingface.co/NYSgpt
What people use it for
- The full USAspending database in SQL without restoring the 177 GB PostgreSQL archive: transactions, awards, sub-awards, recipients and account balances.
- OpenAlex in SQL: 465 million works with authors, institutions, funders and topics.
- Every U.S. patent: PatentsView grants from 1976 and applications from 2001, claims, citations, assignments, examination history and office actions.
- Jobs by census block (LEHD LODES): home-to-work flows and job counts, 2002 to 2024, 15.6 billion rows.
- SEC financial statements, 13F holdings and insider transactions, EPA facility and emissions data, FERC electricity sales, Medicare payments and cost reports, IRS and Form 5500 filings, court opinions.
Connect
Claude (claude.ai or desktop): Settings, Connectors, Add custom connector, and paste https://data.nysgpt.com/mcp?key=YOUR_KEY.
Claude Code:
Cursor, VS Code and other MCP clients (mcp.json):
HTTP:
Tools
HTTP equivalents: GET /databases, GET /tables?database=, GET /table?database=&table=, POST /query.
Example
More in examples/.
Largest datasets
All 132 datasets, with every table and field: https://nysgpt.com/data
Limits and plans
Every call needs a key: add ?key=YOUR_KEY to the MCP address or send the header X-Api-Key.
Notes
- Queries are read-only:
SELECT,WITH,SHOW,DESCRIBE. - Many source files are loaded as published, so numeric columns are often text; use
try_cast. - The data is public records as issued by each publisher. Each dataset page names its publisher. Check the publisher's terms for your use.
- Questions and requests for datasets: open an issue here.
Source: README.md at commit e862d22
Tools
0Version history
1- v2.0.0LatestNov 17, 2026
Score
75% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 daysRequired25 pts
- Auth declaredRequired10 pts
- Tools listedRequired15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts
Declare a license in the repository or the registry entry.