Kalshi

Kalshi API

Kalshi is a CFTC-regulated prediction market exchange where users trade on the outcomes of real-world events.

Check this skill and the official documentation FREQUENTLY. The canonical machine-readable index of all docs lives at:

https://docs.kalshi.com/llms.txt

Always consult llms.txt to discover every available page before building. It is the single source of truth for endpoint references, SDK docs, WebSocket channels, FIX protocol details, and getting-started guides.

Key Concepts (Glossary)

Full glossary: https://docs.kalshi.com/getting_started/terms.md

Base URLs

Note: Despite the elections subdomain, the production URL serves ALL Kalshi markets — economics, weather, tech, entertainment, etc.

Demo environment info: https://docs.kalshi.com/getting_started/demo_env.md

Documentation & References

All detailed examples, request/response schemas, and walkthroughs live in the official docs. Always consult these before building:

Authentication

Kalshi uses RSA-PSS signed requests (not Bearer tokens). Every authenticated request requires three headers:

Important: When signing, use the path without query parameters. For example, sign /trade-api/v2/portfolio/orders even if the request URL is /trade-api/v2/portfolio/orders?limit=5.

Generating API Keys

1. Log in → Account Settings → API Keys section.
2. Click "Create New API Key" to generate an RSA key pair.
3. Store the private key immediately — it cannot be retrieved again.

Full guide: https://docs.kalshi.com/getting_started/api_keys.md

Python Signing Example

import base64, datetime, requests
from cryptography.hazmat.primitives import serialization, hashes
from cryptography.hazmat.primitives.asymmetric import padding

def load_private_key(file_path):
    with open(file_path, "rb") as f:
        return serialization.load_pem_private_key(f.read(), password=None)

def sign_pss(private_key, text: str) -> str:
    sig = private_key.sign(
        text.encode("utf-8"),
        padding.PSS(mgf=padding.MGF1(hashes.SHA256()), salt_length=padding.PSS.DIGEST_LENGTH),
        hashes.SHA256(),
    )
    return base64.b64encode(sig).decode("utf-8")

def kalshi_headers(key_id, private_key, method, path):
    ts = str(int(datetime.datetime.now().timestamp() * 1000))
    path_no_qs = path.split("?")[0]
    sig = sign_pss(private_key, ts + method + path_no_qs)
    return {
        "KALSHI-ACCESS-KEY": key_id,
        "KALSHI-ACCESS-SIGNATURE": sig,
        "KALSHI-ACCESS-TIMESTAMP": ts,
    }

JavaScript Signing Example

const crypto = require("crypto");
const fs = require("fs");

function signPss(privateKeyPem, text) {
  const sign = crypto.createSign("RSA-SHA256");
  sign.update(text);
  sign.end();
  return sign.sign({
    key: privateKeyPem,
    padding: crypto.constants.RSA_PKCS1_PSS_PADDING,
    saltLength: crypto.constants.RSA_PSS_SALTLEN_DIGEST,
  }).toString("base64");
}

function kalshiHeaders(keyId, privateKeyPem, method, path) {
  const ts = Date.now().toString();
  const pathNoQs = path.split("?")[0];
  const sig = signPss(privateKeyPem, ts + method + pathNoQs);
  return {
    "KALSHI-ACCESS-KEY": keyId,
    "KALSHI-ACCESS-SIGNATURE": sig,
    "KALSHI-ACCESS-TIMESTAMP": ts,
  };
}

SDKs

Kalshi provides official SDKs. Prefer these for production integrations:

The old kalshi-python package is deprecated. Migrate to kalshi_python_sync or kalshi_python_async.

SDK overview: https://docs.kalshi.com/sdks/overview.md

For production applications, consider generating your own client from the OpenAPI spec.

Rate Limits

Write-limited endpoints: CreateOrder, CancelOrder, AmendOrder, DecreaseOrder, BatchCreateOrders, BatchCancelOrders. In batch APIs each item counts as 1 transaction (except BatchCancelOrders where each cancel = 0.2 transactions).

Full details: https://docs.kalshi.com/getting_started/rate_limits.md

Pagination

The API uses cursor-based pagination. Responses include a cursor field; pass it as ?cursor={value} on the next request. An empty/null cursor means no more pages. Default page size is 100 (max typically 1000).

Full guide: https://docs.kalshi.com/getting_started/pagination.md

REST API Endpoints Overview

Below is a summary organized by domain. For full request/response schemas, see the linked docs or browse https://docs.kalshi.com/llms.txt.

Markets

Events

Orders (Authenticated)

Portfolio (Authenticated)

Subaccounts (Authenticated)

Order Groups (Authenticated)

Order groups apply a rolling 15-second contracts limit; when hit, all orders in the group are auto-cancelled.

Communications / RFQ (Authenticated)

Historical Data (Authenticated)

Historical data is for archived markets, orders, and fills that have moved past cutoff timestamps.

Historical data guide: https://docs.kalshi.com/getting_started/historical_data.md

Multivariate Collections

Milestones & Structured Targets

Live Data

Search & Discovery

Exchange & Account

Incentive Programs

WebSocket API

A single authenticated WebSocket connection provides real-time streaming. Subscribe to channels by sending JSON commands.

Connection

wss://api.elections.kalshi.com/trade-api/ws/v2      (production)
wss://demo-api.kalshi.co/trade-api/ws/v2            (demo)

Authentication headers (KALSHI-ACCESS-KEY, KALSHI-ACCESS-SIGNATURE, KALSHI-ACCESS-TIMESTAMP) must be included during the handshake. Sign the string {timestamp}GET/trade-api/ws/v2.

Available Channels

*The connection itself always requires authentication, even for public channels.

Subscribe Command

{
  "id": 1,
  "cmd": "subscribe",
  "params": {
    "channels": ["ticker", "orderbook_delta"],
    "market_tickers": ["KXHIGHNY-24JAN01-T60"]
  }
}

Keep-Alive

Kalshi sends Ping frames every 10 seconds with body heartbeat. Clients must respond with Pong frames. Docs: Connection Keep-Alive

Full WebSocket guide: https://docs.kalshi.com/getting_started/quick_start_websockets.md

Orderbook Structure

Kalshi orderbooks return bids only (no asks). In binary markets:

• A YES bid at X¢ = a NO ask at (100−X)¢
• A NO bid at Y¢ = a YES ask at (100−Y)¢

Each entry is [price, quantity] sorted ascending. The best bid is the last element.

Spread calculation:

• Best YES ask = 100 - best_no_bid
• Spread = Best YES ask − Best YES bid

Full guide: https://docs.kalshi.com/getting_started/orderbook_responses.md

FIX Protocol

Kalshi also supports the Financial Information eXchange (FIX) protocol for low-latency trading:

Common Patterns

Fetch All Open Markets for a Series

curl -s "https://api.elections.kalshi.com/trade-api/v2/markets?series_ticker=KXHIGHNY&status=open" | jq '.markets[] | {ticker, title, yes_price, volume}'

Get Orderbook for a Market

curl -s "https://api.elections.kalshi.com/trade-api/v2/markets/KXHIGHNY-24JAN01-T60/orderbook" | jq '.orderbook'

Paginate Through All Results

# First page
curl -s "https://api.elections.kalshi.com/trade-api/v2/markets?limit=100" > page1.json
# Extract cursor for next page
CURSOR=$(jq -r '.cursor' page1.json)
# Next page
curl -s "https://api.elections.kalshi.com/trade-api/v2/markets?limit=100&cursor=$CURSOR" > page2.json

Usage Tips

• Always check llms.txt first: https://docs.kalshi.com/llms.txt has every endpoint and guide.
• Use the demo environment for testing — credentials are separate from production.
• Never hardcode API keys — use environment variables or secure key storage.
• Prices are in cents (1–99). Monetary values in WebSocket position channels are in centi-cents (divide by 10,000 for dollars).
• Sign paths without query parameters. Strip everything after ? before signing.
• Handle pagination — always check for a non-null cursor and loop until exhausted.
• Respect rate limits — implement exponential backoff on 429 responses.
• Combine REST + WebSocket for the most accurate state: use REST for initial snapshots and WebSocket for real-time deltas.
• Orderbook is bids-only — derive asks via the 100−price complement.
• Historical vs. Live data: Check cutoff timestamps to know whether to query live or historical endpoints.
• For market status values, use: unopened, open, closed, settled.
• Batch operations count against per-second write limits per item.

Error Handling

Standard HTTP error codes apply. The API returns JSON error bodies with descriptive messages. Implement retry with backoff for 429 (rate limited) and 5xx (server errors).

WebSocket Error Codes

Full error code table: https://docs.kalshi.com/getting_started/quick_start_websockets.md

Example Project: Alph Bot

Alph Bot is an open-source automated trading bot that demonstrates a production-quality integration of the Kalshi API alongside Shipp for real-time sports data and Claude AI for probability estimation.

How Alph Bot Uses Kalshi

1. Market discovery — Searches for Kalshi event contracts related to a live sports game (e.g., NBA point spreads, totals).
2. Orderbook reading — Fetches orderbook data to determine current market-implied probabilities (prices in cents map directly to implied %).
3. Edge detection — Compares Kalshi market prices against AI-estimated probabilities powered by Claude + Shipp real-time game data.
4. Order execution — Places limit orders when sufficient edge is found, using Kelly Criterion for position sizing.
5. Risk management — Enforces circuit breakers (max daily loss), position size limits, single-market exposure caps, and minimum account balance thresholds.

Alph Bot's Kalshi Configuration

From its .env.example:

ALPH_BOT_KALSHI_API_KEY_ID=abc123
ALPH_BOT_KALSHI_PRIVATE_KEY_PATH=./keys/kalshi-private.pem

# Strategy
ALPH_BOT_MIN_EDGE_PCT=5
ALPH_BOT_MIN_CONFIDENCE=medium
ALPH_BOT_KELLY_FRACTION=0.25

# Risk controls
ALPH_BOT_MAX_TOTAL_EXPOSURE_USD=10000
ALPH_BOT_MAX_POSITION_SIZE_USD=1000
ALPH_BOT_MAX_SINGLE_MARKET_PERCENT=20
ALPH_BOT_MAX_DAILY_LOSS_USD=500
ALPH_BOT_MAX_DAILY_TRADES=50
ALPH_BOT_MIN_ACCOUNT_BALANCE_USD=100

Try It

git clone https://gitlab.com/outsharp/shipp/alph-bot.git
cd alph-bot
cp .env.example .env
# Add your Kalshi, Shipp, and Anthropic API keys

yarn migrate

# Find a game to trade on
./index.ts available-games --sport NBA

# Run value betting in demo mode (uses Kalshi demo environment)
./index.ts value-bet -d --game <GAME_ID>

# Run in paper mode (no real orders executed)
./index.ts value-bet --paper --game <GAME_ID>

Warning: Trading on Kalshi involves real money when not in demo/paper mode. Always start with a demo account.

See the Alph Bot README for full setup instructions.

Versioning

The current API version is v2 under /trade-api/v2. Monitor the API Changelog for updates. SDK versions align with the OpenAPI spec and are generally published weekly.