Data

MCP server

The Kalshi Eye MCP server lets AI assistants such as Claude and Cursor search the archive. Ask about Kalshi, Polymarket and Limitless markets, how a market traded, what was recorded, and which files hold it. Searching needs no account.

Install

It runs on your machine through uv's uvx, which fetches the kalshieye-mcp package from PyPI.

Claude Code

claude mcp add kalshieye -- uvx kalshieye-mcp

Claude Desktop, Cursor and other clients

{
  "mcpServers": {
    "kalshieye": {
      "command": "uvx",
      "args": ["kalshieye-mcp"]
    }
  }
}

Put that in claude_desktop_config.json (Claude Desktop: Settings, Developer, Edit config) or .cursor/mcp.json. Without uv, run pip install kalshieye-mcp and use kalshieye-mcp as the command.

To list and download files too, add your key from API to that entry as "env": { "KALSHIEYE_API_KEY": "clab_…" }. A missing or placeholder key is ignored, and a rejected key only affects the two file tools.

Tools

Tool	What it does	API key
`search`	Find series, events and markets by keyword, venue, category or status	No
`get_series`	One series: what it is, how often it runs, markets in the catalog, and days of data files	No
`list_markets`	A series' most recent markets, newest first, with status and result	No
`get_market`	One market's terms, status, result, and how much of it was recorded	No
`market_price_history`	A short recorded YES bid/ask path, up to about 40 points	No
`list_datasets`	Every dataset: what it holds, its columns, plan and days available	No
`dataset_series`	Which series have files in a dataset, with days, rows and size	No
`recording_status`	Whether recording is running, and the recorded share of recent days by venue	No
`list_files`	Daily Parquet files for a dataset, date range and series, page by page	Yes
`download_links`	Signed download URLs for up to 100 of those files (counts toward the monthly allowance)	Yes

The last two use an API key from API, set as KALSHIEYE_API_KEY, and work for the datasets your plan includes. Everything else reads the public API (which calls are public).

Things to ask

Which bitcoin markets does Kalshi Eye record, and since when?
What were the last few hours of the Bills vs Rams market on October 12 like?
How much of the Fed decision series is in the order book files?
Is the recorder running, and how complete was last week?
List the trades files for KXBTC15M on October 13 and give me download links.

Prices come back as dollars between 0 and 1 and times in UTC. For analysis, the assistant can hand you file links; read them with the code in the backtest guide or the data dictionary.

Source and registry

Code: KalshiEye/kalshieye-mcp on GitHub, MIT licensed.
Package: kalshieye-mcp on PyPI.
MCP Registry: io.github.KalshiEye/kalshieye-mcp.