Kalshi MCP Server
Give Claude access to Kalshi prediction markets โ browse live odds, analyze price history, and trade YES/NO contracts.
README
Kalshi MCP Server
Give Claude access to Kalshi prediction markets โ browse live odds, analyze price history, and trade YES/NO contracts.
Built on the Model Context Protocol (MCP), this server connects Claude (and any MCP-compatible AI) directly to the Kalshi Trade API v2.
First MCP server for Kalshi. Ask Claude "What are the current odds on the Fed rate decision?" and get a live answer.
Features
| Tool | Description | Tier |
|---|---|---|
get_markets |
Browse open/settled markets with filters | ๐ Free |
get_market_details |
Live odds, volume, order book for a market | ๐ Free |
get_market_price_history |
OHLC candlestick data (1-min, hourly, daily) | ๐ Free |
get_portfolio |
Account balance and open positions | ๐ Pro |
place_order |
Buy YES/NO contracts (dry-run confirmation) | ๐ Pro |
get_order_history |
Fills, settlements, open orders | ๐ Pro |
Quick Start
1. Install
git clone https://github.com/shadowfax-mitch/kalshi-mcp-server
cd kalshi-mcp-server
pip install -r requirements.txt
Or install as a package:
pip install kalshi-mcp-server
2. Configure Claude Desktop
Add to your claude_desktop_config.json:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Free Tier (read-only, no credentials needed)
{
"mcpServers": {
"kalshi": {
"command": "python",
"args": ["-m", "kalshi_mcp.server"],
"cwd": "/path/to/kalshi-mcp-server"
}
}
}
Pro Tier (full trading access)
{
"mcpServers": {
"kalshi": {
"command": "python",
"args": ["-m", "kalshi_mcp.server"],
"cwd": "/path/to/kalshi-mcp-server",
"env": {
"KALSHI_API_KEY": "your-api-key-uuid",
"KALSHI_PRIVATE_KEY_PATH": "/absolute/path/to/private_key.pem"
}
}
}
}
Pro Tier (with package install / pipx)
{
"mcpServers": {
"kalshi": {
"command": "kalshi-mcp",
"env": {
"KALSHI_API_KEY": "your-api-key-uuid",
"KALSHI_PRIVATE_KEY_PATH": "/absolute/path/to/private_key.pem"
}
}
}
}
3. Restart Claude Desktop
After saving the config, restart Claude Desktop. You'll see a ๐จ tools icon โ click it to verify Kalshi tools are loaded.
Getting Kalshi API Credentials (Pro Tier)
- Go to https://kalshi.com/account/settings
- Navigate to API โ Create API Key
- Download your RSA private key PEM file
- Copy your API key UUID
- Set both as environment variables (see config above)
Security: Never commit your private key or API key to version control.
Store them securely โ they provide full trading access to your account.
Usage Examples
Once configured, talk to Claude naturally:
Browse Markets
"Show me open BTC prediction markets"
โ get_markets(series_ticker="KXBTC15M")
"What are the most liquid markets right now?"
โ get_markets(status="open", limit=20)
"Show me recently settled markets"
โ get_markets(status="settled", limit=10)
Market Research
"Get the full details for KXBTC15M-24NOV1913-T10000 including the order book"
โ get_market_details("KXBTC15M-24NOV1913-T10000")
"Show me the hourly price chart for this market over the last 48 hours"
โ get_market_price_history("KXBTC15M-...", period_interval=60, lookback_hours=48)
"Get 1-minute candles for the last hour"
โ get_market_price_history("...", period_interval=1, lookback_hours=1)
Portfolio (Pro)
"What's my current balance and positions?"
โ get_portfolio()
"Show me my recent trades and P&L"
โ get_order_history(include_fills=True, include_settlements=True)
Trading (Pro)
Claude automatically uses a dry-run confirmation flow:
"Buy 10 YES contracts on KXBTC15M-... at 65 cents"
Claude:
1. Calls place_order(..., dry_run=True) โ preview
2. Shows you:
๐ ORDER PREVIEW (not submitted):
Ticker: KXBTC15M-24NOV1913-T10000
Side: YES
Contracts: 10
Limit price: 65ยข ($0.65)
Estimated cost: $6.50
Max payout: $10.00
Max profit: $3.50
3. Asks: "Shall I submit this order?"
4. On confirmation: calls place_order(..., dry_run=False)
Tool Reference
get_markets
List Kalshi prediction markets.
| Parameter | Type | Default | Description |
|---|---|---|---|
status |
str | "open" |
"open", "closed", or "settled" |
series_ticker |
str | Filter by series (e.g. "KXBTC15M", "NASDAQ100D") |
|
event_ticker |
str | Filter by event | |
limit |
int | 20 |
Results per page (max 200) |
cursor |
str | Pagination cursor |
get_market_details
Full detail for a single market including live order book.
| Parameter | Type | Default | Description |
|---|---|---|---|
ticker |
str | required | Market ticker |
include_orderbook |
bool | true |
Fetch live order book |
get_market_price_history
OHLC candlestick price history.
| Parameter | Type | Default | Description |
|---|---|---|---|
ticker |
str | required | Market ticker |
period_interval |
int | 60 |
Candle size in minutes: 1, 60, or 1440 |
lookback_hours |
int | 24 |
Hours of history (1โ168) |
get_portfolio (Pro)
Account balance and open positions. No parameters.
place_order (Pro)
Place a limit buy order.
| Parameter | Type | Default | Description |
|---|---|---|---|
ticker |
str | required | Market ticker |
side |
str | required | "yes" or "no" |
count |
int | required | Number of contracts |
price_cents |
int | required | Limit price in cents (1โ99) |
dry_run |
bool | true |
Preview without submitting (always start here) |
Pricing guide:
- Each contract pays $1.00 if you're correct, $0.00 if wrong
price_cents=65means you pay 65ยข and profit 35ยข if correctprice_centsreflects the market's implied probability (65ยข โ 65% chance)
get_order_history (Pro)
Recent trading activity.
| Parameter | Type | Default | Description |
|---|---|---|---|
include_fills |
bool | true |
Recent executed trades |
include_settlements |
bool | true |
Settled contracts with P&L |
include_open_orders |
bool | true |
Resting limit orders |
ticker |
str | Filter to specific market | |
limit |
int | 25 |
Records per section |
Architecture
Claude Desktop
โ
โ MCP (stdio)
โผ
kalshi_mcp/server.py โ FastMCP server, tool definitions, formatting
โ
โผ
kalshi_mcp/client.py โ Kalshi REST API client (httpx)
โ
โผ
kalshi_mcp/auth.py โ RSA-PSS signing (Kalshi API v2 auth scheme)
โ
โผ
https://api.elections.kalshi.com/trade-api/v2
Auth scheme (RSA-PSS):
Signature = RSA-PSS(
message = f"{unix_timestamp}{METHOD}{/trade-api/v2/endpoint}",
hash = SHA-256,
mgf = MGF1(SHA-256),
salt_len = 32
)
Headers:
KALSHI-ACCESS-KEY = api_key_uuid
KALSHI-ACCESS-SIGNATURE = base64(signature)
KALSHI-ACCESS-TIMESTAMP = unix_timestamp_seconds
Development
# Install with dev dependencies
pip install -e ".[dev]"
# Run tests
pytest tests/ -v
# Type check
mypy kalshi_mcp/
# Lint
ruff check kalshi_mcp/
# Test the server manually (stdio transport)
python -m kalshi_mcp.server
Project Structure
kalshi-mcp-server/
โโโ kalshi_mcp/
โ โโโ __init__.py # Package metadata
โ โโโ server.py # MCP server & tool definitions
โ โโโ client.py # Kalshi API client
โ โโโ auth.py # RSA-PSS authentication
โโโ tests/
โ โโโ test_client.py # Unit tests
โโโ README.md
โโโ requirements.txt
โโโ pyproject.toml
โโโ .env.example
Pricing & Tiers
๐ Free Tier
- Market browsing โ search and filter thousands of markets
- Live odds โ real-time YES/NO prices and order books
- Price history โ OHLC candlestick data for any market
- No API credentials required
๐ Pro Tier
Everything in Free, plus:
- Portfolio view โ balance, positions, unrealized P&L
- Order placement โ buy YES/NO contracts with confirmation flow
- Order history โ fills, settlements, open orders
Available at: PaidMCP | GitHub Sponsors
Security
- Never trade with funds you can't afford to lose. Prediction markets carry real financial risk.
- Store your API key and private key securely. Never commit them to version control.
- The
place_ordertool defaults todry_run=Trueโ orders require explicit confirmation. - Consider setting position size limits at the Kalshi account level as an extra safeguard.
Troubleshooting
"Auth init failed (free-tier only)"
โ Check that KALSHI_API_KEY is a valid UUID and KALSHI_PRIVATE_KEY_PATH points to your PEM file.
"Kalshi API 401"
โ Your key may be expired or revoked. Regenerate at kalshi.com/account/settings.
"Kalshi API 403"
โ You may be trying a pro feature without valid credentials, or your account may be restricted.
Tools not showing in Claude Desktop
โ Verify the cwd path in your config. Restart Claude Desktop after any config change.
"period_interval must be 1, 60, or 1440"
โ Kalshi only supports these candle sizes. Use 1 (minute), 60 (hourly), or 1440 (daily).
Contributing
Pull requests welcome! Please open an issue first for significant changes.
git clone https://github.com/shadowfax-mitch/kalshi-mcp-server
cd kalshi-mcp-server
pip install -e ".[dev]"
pytest tests/
License
MIT License โ see LICENSE
Disclaimer
This is an unofficial third-party tool. It is not affiliated with, endorsed by, or sponsored by Kalshi Inc. Use of the Kalshi API is subject to Kalshi's Terms of Service. Trading prediction markets involves financial risk.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.