Quant
БесплатноНе проверенPrediction-market quant tools for AI agents: EV, Kelly, arbitrage, edge for Kalshi & Polymarket.
Описание
Prediction-market quant tools for AI agents: EV, Kelly, arbitrage, edge for Kalshi & Polymarket.
README
A hosted Model Context Protocol server that gives AI agents institutional-grade quant tools for Kalshi and Polymarket prediction markets — expected value, Kelly sizing, Bayesian updating, probability conversion, cross-platform arbitrage, and live edge signals.
- Endpoint (Streamable HTTP):
https://predictionmarketspicks.com/api/mcp/mcp - Registry name:
com.predictionmarketspicks/quant(Model Context Protocol registry) - Docs / landing page: https://predictionmarketspicks.com/mcp
- Manifest: server.json · also served at
https://predictionmarketspicks.com/.well-known/mcp/server.json - Type: Cloud service ☁️ · TypeScript 📇
This repo is the public home and documentation for the hosted server. The server is live — no install, no build. Point any MCP client at the endpoint above.
It also ships a local, self-contained build of the six free calculators (src/) — a stdio MCP server with no network access that runs in any sandbox. See Run the free tier locally. The four Pro tools read live PMP edge engines and are only available on the hosted endpoint.
Tools
The free tier is a set of stateless quant calculators. Pro tools read PredictionMarketsPicks' live edge engines and require an API key.
| Tool | Tier | What it does |
|---|---|---|
calculate_ev |
Free | Expected-value edge on a contract from market price + your probability; returns edge % and a BUY / SELL / SKIP read. |
kelly_size |
Free | Optimal Kelly position size (full / half / quarter / eighth) from win probability, price, and bankroll, with a risk rating. |
bayes_update |
Free | Update a prior with one or more pieces of evidence via Bayes' theorem; returns the posterior and the per-step chain. |
convert_probability |
Free | Convert between implied probability, American odds, and decimal odds (American odds carry no commas). |
base_rate_gap |
Free | Compare a market price to the historical base rate for a class of events; returns the gap in points + sample-size quality. |
combo_edge |
Free | Grade a same-game multi-leg combo: EV %, fair vs offered odds, and a negative-correlation-trap flag. |
find_arbitrage |
Pro | Cross-platform price gaps between Kalshi and Polymarket on the same sports contract (NBA, NHL, MLB, World Cup). |
market_pulse |
Pro | US macro-health composite (0–100) and regime, plus six category scores. |
commodity_edge |
Pro | Largest model edge on a Kalshi weekly-silver or twice-daily bitcoin strike, as a trade ticket (side, price, criterion, edge, tier, ¼-Kelly). |
scan_mispricings |
Pro | Polymarket contracts trading away from the PMP model, with direction, edge in points, and quarter-Kelly sizing. |
All tool descriptions and outputs use prediction-market terminology (trader / position / contract / market analysis).
Connect
Claude Code
claude mcp add --transport http predictionmarketspicks https://predictionmarketspicks.com/api/mcp/mcp
Claude.ai, ChatGPT, or Cursor — add a custom connector / MCP server with the URL:
https://predictionmarketspicks.com/api/mcp/mcp
The free calculators work with no key. Pro tools require a PredictionMarketsPicks API key — see https://predictionmarketspicks.com/mcp.
Run the free tier locally
The six free calculators run entirely offline as a stdio MCP server — no key, no network. Useful for air-gapped agents, testing, or sandboxed hosts.
npm install
npm start # stdio MCP server: 6 free tools
npm run smoke # end-to-end self-test
Or with Docker:
docker build -t pmp-mcp-quant .
docker run --rm -i pmp-mcp-quant
Point a stdio MCP client at node src/index.js (or the container). For the full ten-tool experience including the live Pro edge engines, use the hosted endpoint above.
About
Built by PredictionMarketsPicks — independent quant tools and edge analysis for Kalshi and Polymarket, published by The 7 Oracles.
License
MIT — see LICENSE. The hosted service and its live data are operated by PredictionMarketsPicks; this repository covers the server's public interface and documentation.
Установка Quant
У этого сервера нет опубликованного пакета — он собирается из исходников. Открой репозиторий и следуй инструкции в README.
▸ github.com/predictionmarketspicks/mcpFAQ
Quant MCP бесплатный?
Да, Quant MCP бесплатный — установка в пару кликов через Unyly без оплаты.
Нужен ли API-ключ для Quant?
Нет, Quant работает без API-ключей и переменных окружения.
Quant — hosted или self-hosted?
Доступен hosted-вариант: Unyly запускает сервер в облаке, локальная установка не обязательна.
Как установить Quant в Claude Desktop, Claude Code или Cursor?
Открой Quant на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.
Похожие MCP
Fetch
Web content fetching and conversion for efficient LLM usage.
AWS KB Retrieval
Retrieval from AWS Knowledge Base using Bedrock Agent Runtime.
автор: modelcontextprotocolSpring AI MCP Server
Provides auto-configuration for setting up an MCP server in Spring Boot applications.
llm-analysis-assistant
A very streamlined mcp client that supports calling and monitoring stdio/sse/streamableHttp, and can also view request responses through the /logs page. It also
автор: xuzexin-hzCompare Quant with
Не уверен что выбрать?
Найди свой стек за 60 секунд
Автор?
Embed-бейдж для README
Похожее
Все в категории ai
