mcpField
A
assymetrix

assymetrix

Local
github.com/assymetrixOther0Updated today

About assymetrix

assymetrix provides read-only prediction-market data from Polymarket, Kalshi, and Limitless for AI agents. It includes an MCP server and an Agent Skill using the same API; neither places orders nor handles funds.

Key features

  1. 01Search markets by question text
  2. 02Retrieve current odds and price details
  3. 03Read OHLCV history or recent trades
  4. 04Access Polymarket, Kalshi, and Limitless data
  5. 05Use an MCP server or Agent Skill

Use cases

  1. 01Find markets related to a question
  2. 02Compare a market's prices and liquidity
  3. 03Review price history or recent trades
  4. 04Give an agent prediction-market data

MCP support

The MCP server exposes three read-only tools backed by the Assymetrix Data API: market search, current prices, and market history. It requires an Assymetrix API key.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"assymetrix": {
"command": "npx",
"args": [
"-y",
"@assymetrix/mcp-server"
],
"env": {
"ASSYMETRIX_API_KEY": "<ASSYMETRIX_API_KEY>"
}
}
}
}
Run commandnpx -y @assymetrix/mcp-server
Environment variables to set
ASSYMETRIX_API_KEYSecretYour Assymetrix Data API key (starts with asy_). A free key works.

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@assymetrix/mcp-server0.1.1

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

Assymetrix agent tools

Prediction market data from Polymarket, Kalshi and Limitless for AI agents, from the Assymetrix Data API:

Everything here only reads market data. Nothing places orders, moves money or links wallets, and nothing here is financial advice. You use your own Assymetrix API key, and a free key works.

Get an API key

  1. Create a free account at https://dashboard.assymetrix.com/signup?utm_source=github&utm_campaign=agent-tools
  2. Open API Keys and create a key. It is shown once and starts with asy_.
  3. Keep it in an environment variable, for example in your shell profile: export ASSYMETRIX_API_KEY=asy_...

The Free plan includes 10,000 requests per rolling 30 days, 1 request per second and 30 days of history. Each tool call is one request. Paid plans: https://dashboard.assymetrix.com/billing

Install

Recommended: the Claude Code plugin (MCP server and skill)
/plugin marketplace add assymetrix/assymetrix-agent-tools
/plugin install assymetrix@assymetrix

Claude Code asks for your API key when you enable the plugin and keeps it in your system's secure store, so the key never sits in a config file or the chat.

Other clients

These read the key from an environment variable or their own config file. Keep the key in your shell profile (export ASSYMETRIX_API_KEY=asy_...) rather than typing it into a command.

Claude Code without the plugin (--transport goes before the name):

claude mcp add --env ASSYMETRIX_API_KEY="$ASSYMETRIX_API_KEY" --transport stdio assymetrix -- npx -y @assymetrix/mcp-server

Codex:

codex mcp add assymetrix --env ASSYMETRIX_API_KEY="$ASSYMETRIX_API_KEY" -- npx -y @assymetrix/mcp-server

Cursor (~/.cursor/mcp.json or .cursor/mcp.json):

{
  "mcpServers": {
    "assymetrix": {
      "command": "npx",
      "args": ["-y", "@assymetrix/mcp-server"],
      "env": { "ASSYMETRIX_API_KEY": "${env:ASSYMETRIX_API_KEY}" }
    }
  }
}

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "assymetrix": {
      "command": "npx",
      "args": ["-y", "@assymetrix/mcp-server"],
      "env": { "ASSYMETRIX_API_KEY": "asy_..." }
    }
  }
}

VS Code (.vscode/mcp.json, which asks for the key and hides it):

{
  "inputs": [
    {
      "type": "promptString",
      "id": "assymetrix-key",
      "description": "Assymetrix API key",
      "password": true
    }
  ],
  "servers": {
    "assymetrix": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@assymetrix/mcp-server"],
      "env": { "ASSYMETRIX_API_KEY": "${input:assymetrix-key}" }
    }
  }
}
Skill only (Claude Code, Codex, Cursor and other agents)
npx skills add assymetrix/assymetrix-agent-tools

The skill reads ASSYMETRIX_API_KEY from the agent's environment and never prints it.

Tools

ToolWhat it returns
search_marketsMarkets whose question contains your text, sorted by 24-hour volume: id, venue, last price, best bid and ask, end date.
get_market_pricesOne market's current odds: last price, bid, ask, spread, price changes, each outcome's price, volume and liquidity.
get_market_historyOHLCV candles for a time range (the interval is picked to fit), or the market's trades, newest first.

Prices are probabilities from 0 to 1. How far back history goes depends on your plan; when your plan cuts a range, the result says so.

Configuration

VariableDefaultMeaning
ASSYMETRIX_API_KEYnoneYour API key. Without it the server starts and explains how to get one.
ASSYMETRIX_API_BASE_URLhttps://api.assymetrix.com/api/v1Only for testing. Must be https (plain http only for localhost).
ASSYMETRIX_TIMEOUT_MS15000Timeout for one API request, 1000 to 60000.
ASSYMETRIX_LOG_LEVELwarnerror, warn, info or debug. Logs go to stderr.

Privacy

  • The MCP server talks only to the Assymetrix Data API, sending your key in the x-api-key header, a User-Agent with its version, and a random request id that support can use to find a call.
  • It collects no telemetry, writes nothing to disk, and never logs or returns your key.
  • The skill runs curl against the same API from your own machine.
  • Privacy policy: https://www.assymetrix.com/privacy-policy. Support: [email protected].

Development

cd packages/mcp-server
pnpm install
pnpm run check      # type check, build, then the tests (no network)
node ../../skills/validate-skill.cjs ../../skills/prediction-market-data            # skill format checks
node ../../skills/validate-skill.cjs ../../skills/prediction-market-data --release  # also fails on open TODO markers

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.