Dataville — Claude Code MCP server: what it does, how to install
Links AI agents to ten public data sources through one interface. Useful for pulling SEC filings, arXiv research, and US Census numbers directly into chats.
claude mcp add --transport http dataville-mcp https://api.dataville.com/mcpWhat Dataville does
The server exposes ten external datasets, including arXiv, SEC EDGAR, US Census, PyPI, OpenAlex, and Project Gutenberg. MCP clients interact with them through four read-only operations: checking source catalogs, inspecting table schemas, running text searches, and sending SQL queries.
The query_dataville tool handles SQL SELECT statements with filtering, joins, and pagination capped at 1,000 rows per call. Datasets for census and news are excluded from SQL queries. Dataville fetches both live on each search.
You can link directly to the remote HTTP endpoint without local packages. Running the local stdio version instead requires Node.js LTS and a valid DATAVILLE_API_KEY.
What it helps with
- inspect corporate reports and financial disclosures from SEC EDGAR
- locate preprints on arXiv and extract paper summaries
- retrieve county demographic metrics from the US Census
- look up package licenses and latest releases on PyPI
Example prompts for the agent
- Check the latest SEC EDGAR filing for Apple and summarize revenue details.
- Search arXiv for retrieval-augmented generation papers and show abstracts.
- Find the median household income for Travis County, Texas in US Census data.
How to install Dataville
1Platform CLI: remote server
claude mcp add --transport http dataville-mcp https://api.dataville.com/mcp2Platform CLI
claude mcp add --env DATAVILLE_API_KEY=<DATAVILLE_API_KEY> --env DATAVILLE_API_BASE_URL=<DATAVILLE_API_BASE_URL> --transport stdio dataville -- npx -y @dataville/dataville-mcpReplace <…> with your own values
1Platform CLI: remote server
codex mcp add dataville --url https://api.dataville.com/mcp2Platform CLI
codex mcp add dataville --env DATAVILLE_API_KEY=<DATAVILLE_API_KEY> --env DATAVILLE_API_BASE_URL=<DATAVILLE_API_BASE_URL> -- npx -y @dataville/dataville-mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"dataville": {
"url": "https://api.dataville.com/mcp"
}
}
}{
"mcpServers": {
"dataville": {
"command": "npx",
"args": [
"-y",
"@dataville/dataville-mcp"
],
"env": {
"DATAVILLE_API_KEY": "<DATAVILLE_API_KEY>",
"DATAVILLE_API_BASE_URL": "<DATAVILLE_API_BASE_URL>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"dataville": {
"type": "http",
"url": "https://api.dataville.com/mcp"
}
}
}{
"servers": {
"dataville": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@dataville/dataville-mcp"
],
"env": {
"DATAVILLE_API_KEY": "<DATAVILLE_API_KEY>",
"DATAVILLE_API_BASE_URL": "<DATAVILLE_API_BASE_URL>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http dataville https://api.dataville.com/mcp2Platform CLI
gemini mcp add -e DATAVILLE_API_KEY=<DATAVILLE_API_KEY> -e DATAVILLE_API_BASE_URL=<DATAVILLE_API_BASE_URL> dataville npx -y @dataville/dataville-mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Dataville
How do you install Dataville in Claude Code?
For hosted use without local files, run claude mcp add --transport http dataville https://api.dataville.com/mcp. To run it locally via Node.js, run `claude mcp add --env DATAVILLE_API_KEY=<DATAVILLE_API_KEY> --env DATAVILLE_API_BASE_URL=<DATAVILLE_API_BASE_URL> --transport stdio dataville -- npx -y @dataville/dataville-mcp`.
What does the Dataville MCP server do?
It lets MCP clients query ten public repositories, such as arXiv, SEC filings, US Census, and PyPI. The agent can search by terms or run read-only SQL commands on supported tables.
Can all Dataville datasets be queried with SQL?
No. The query_dataville tool supports stored tables with a 1,000 row ceiling, but census and news bypass SQL and only support live keyword searches.
