telekash-mcp-server
Prediction market probability oracle for AI agents. 26 tools across 500+ live markets from Kalshi and Polymarket. Cross-source arbitrage detection, structured TPF signals, Kelly Criterion sizing, agent performance tracking, and webhook alerts.
README
TeleKash MCP Server
Prediction market probability oracle for AI agents.
What is this?
TeleKash is the probability oracle for the agent economy. Real-time access to 4,500+ live prediction markets from Kalshi (CFTC-regulated) and Polymarket. Structured trading signals, cross-source arbitrage detection, noise filtering, and smart trade routing across crypto, politics, sports, entertainment, finance, weather, tech, and science.
What makes it different:
- Multi-source — Kalshi + Polymarket + native pools, not single-exchange
- Self-calibrating — Platt scaling with daily recalibration cycles and Brier score tracking
- Universal Payment Layer — agents pay per-call with USDC (Base/Polygon/Solana via x402), fiat (Stripe), or TON
- 28 tools — from free probabilities to broker execution with 1% commission
Quick Install
npx telekash-mcp-server
Claude Code
claude mcp add telekash-oracle npx telekash-mcp-server
Claude Desktop / Cursor
Add to your MCP configuration file:
{
"mcpServers": {
"telekash-oracle": {
"command": "npx",
"args": ["telekash-mcp-server"]
}
}
}
Tools
28 tools across 4 categories. All tools work without an API key on the Free tier (100 queries/day). 4,500+ live markets.
Intelligence
| Tool | What it does | Tier |
|---|---|---|
get_probability |
Real-time YES/NO probability with volume-weighted confidence score | Free |
list_markets |
Browse markets by category, source, jurisdiction, with sorting | Free |
search_markets |
Full-text search across 4,500+ markets | Free |
get_history |
Historical probability changes over 1h, 24h, 7d, 30d | Free |
get_sentiment |
AI sentiment analysis with conviction, momentum, and noise filter | Free |
get_market_stats |
Aggregate statistics across all markets and sources | Free |
get_trending |
Markets with the biggest probability swings | Free |
compare_sources |
Kalshi vs Polymarket side-by-side odds comparison | Calibration |
Analytics
| Tool | What it does | Tier |
|---|---|---|
detect_arbitrage |
Cross-source arbitrage opportunities with buy/sell signals | Calibration |
get_signal |
Structured TPF signal: probability + sentiment + noise filter + verdict | Calibration |
get_divergences |
Markets where sources disagree most (STRONG/MODERATE/WEAK) | Calibration |
get_edge |
Kelly Criterion sizing, expected value, and risk classification | Calibration |
track_prediction |
Record predictions for accuracy tracking | Calibration |
get_performance |
Brier score, calibration curve, edge-vs-market analysis | Calibration |
Trading
| Tool | What it does | Tier |
|---|---|---|
execute_trade |
Route trades to Kalshi, Polymarket, or native parimutuel pools | Edge |
get_order_status |
Check fill status, price, and commission on broker orders | Edge |
cancel_order |
Cancel pending or submitted broker orders | Edge |
get_pool_status |
Native pool composition, participant counts, implied odds | Edge |
get_agent_balance |
Agent balance, P&L, win rate, and pool position count | Edge |
get_resolution_status |
Multi-source resolution verification and confidence levels | Edge |
create_market |
Create custom binary prediction markets | Edge |
export_data |
Bulk export: probability history, resolutions, catalogs, arbitrage | Edge |
Admin
| Tool | What it does | Tier |
|---|---|---|
generate_api_key |
Generate a free API key (no signup required) | Free |
get_usage |
Check current API usage, rate limits, and tier status | Free |
register_alert |
Webhook alerts for probability crosses, mispricings, volume spikes | Edge |
list_alerts |
List active webhook alerts with delivery stats | Edge |
delete_alert |
Delete a webhook alert | Edge |
get_health |
System health: connectivity, data freshness, broker status | Free |
Pricing
Per-query pricing. No subscriptions. Free tier requires no API key.
| Tier | Cost | Queries/Day | What you get |
|---|---|---|---|
| Free | $0 | 100 | Probabilities, search, trending, sentiment, stats, health |
| Calibration | $0.01/query | 1,000 | + Cross-source comparison, arbitrage, signals, divergence, Kelly sizing, performance tracking |
| Edge | $0.05/query | Unlimited | + Broker trading (1% commission), native pools (5% fee), webhooks, data export, market creation |
Revenue model
| Stream | Rate | Description |
|---|---|---|
| Intelligence queries | $0 - $0.05/query | Real-time probability, sentiment, signals |
| Broker trades | 1% commission | Best-price execution routed to Kalshi or Polymarket |
| Native pool trades | 5% fee at resolution | Parimutuel pools alongside Telegram users |
Universal Payment Layer
Agents can also pay per-call without an API key or tier — attach payment proof directly in tool arguments.
| Rail | Network | Asset | Fee |
|---|---|---|---|
| x402 (Coinbase) | Base, Polygon, Solana | USDC | Free (0%) |
| Stripe MPP | Fiat | USD (cards, bank) | 2.9% |
| TON | TON mainnet | TON | Free (0%) |
// Agent pays per-call — no subscription needed
const result = await client.callTool({
name: "get_signal",
arguments: {
query: "Bitcoin 200K",
x402_payment: {
tx_hash: "0xabc...",
network: "base",
rail: "x402",
},
},
});
Get an API key
# Via the MCP server itself (free tier, no signup)
# Call the generate_api_key tool
# Or via Telegram
# Message @TeleKashBot with /apikey
Example
Connect to the server and query a market probability:
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";
const transport = new StdioClientTransport({
command: "npx",
args: ["telekash-mcp-server"],
});
const client = new Client({ name: "my-agent", version: "1.0.0" });
await client.connect(transport);
// Get probability for a market
const result = await client.callTool({
name: "get_probability",
arguments: { query: "Will Bitcoin hit $200K by end of 2026?" },
});
console.log(result.content);
// => { probability: { yes: 34, no: 66 }, confidence: { grade: "HIGH", score: 82 }, ... }
More examples in the examples/ directory:
quick-start.ts-- Connect, search, get probabilitiesarbitrage-scanner.ts-- Find cross-source mispricingsportfolio-scanner.ts-- Scan markets and build a ranked portfoliomarket-monitor.ts-- Watch a market and log probability changes
Environment Variables
The server works without any credentials (returns demo data). For live market access:
SUPABASE_URL=https://rrkjtdnxkscukexbsrue.supabase.co
SUPABASE_ANON_KEY=your-anon-key
TELEKASH_API_KEY=your-api-key # Optional: enables paid tiers
TELEKASH_PAYMENT_ADDRESS=0x... # Optional: EVM wallet for x402 USDC payments
TELEKASH_TON_ADDRESS=UQ... # Optional: TON wallet for TON payments
STRIPE_SECRET_KEY=sk_... # Optional: Stripe for fiat payments
X402_FACILITATOR_URL=https://x402.org/facilitator # Default: Coinbase facilitator
Data Sources
| Source | Type | Sync Frequency |
|---|---|---|
| Kalshi | CFTC-regulated US exchange | Every 15 minutes |
| Polymarket | Crypto-native exchange | Every hour |
All sources are aggregated, deduplicated, and categorized automatically across 8 categories: crypto, politics, sports, entertainment, finance, weather, tech, science.
Links
License
MIT
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.