convez.ai
MCP ServerFinanceiroIntermediário
aland4747

Awesome Prediction Markets

aland4747/awesome-prediction-markets

APIs e datasets de mercados de previsão para builders e agentes IA

Ver no GitHub
InicialPluginScore · de 0 a 100

Curadoria abrangente de APIs, datasets, ferramentas e pesquisa para construir aplicações com dados de mercados de previsão. Fornece acesso programático a plataformas como Kalshi, Polymarket e Metaculus, ideal para desenvolvedores e agentes de IA que precisam de probabilidades calibradas em tempo real para eventos globais.

Instalação

$npx convez-ai aland4747/awesome-prediction-markets

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
9
Acompanhando
1
Saúde
42%
Último commit
há 1 mês
Versão
—
Issues abertas
8
Contribuidores
2

Casos de uso

  • Integrar dados de probabilidades de mercados em aplicações e agentes IA
  • Treinar modelos com datasets históricos de previsões calibradas
  • Construir ferramentas de análise e detecção de contagion entre mercados
  • Automatizar pesquisa de oportunidades de trading baseada em dados públicos

Veredito

Recurso sólido e bem-organizado para quem precisa trabalhar com dados de mercados de previsão em escala. Documentação clara com exemplos cURL prontos e comparação técnica entre plataformas; foco prático em APIs reais, não abstrações teóricas.

README

A tradução para o português deste README ainda está na fila.

Awesome Prediction Markets Awesome

APIs, datasets, tools, and research for building on prediction markets. For developers and AI agents, not dashboards.

Prediction markets are the best real-time source of calibrated probability for world events. This list focuses on what you need to build with that data — APIs to fetch it, tools to process it, datasets to train on, and research to understand it.

Pull requests welcome. See contributing guidelines.

Open In Colab — Interactive notebook: world state, market search, trade ideas, contagion detection, and more.

Quick Start — Get a Price in 30 Seconds

# Kalshi — what are the markets?
curl -s "https://trading-api.kalshi.co/trade-api/v2/markets?limit=3" | jq '.markets[] | {title, yes_ask, volume}'

# Polymarket — current markets
curl -s "https://clob.polymarket.com/markets?limit=3" | jq '.[] | {question, tokens[0].price}'

# SimpleFunctions — scan any topic, no auth
curl -s "https://simplefunctions.dev/api/public/scan?q=oil"

# SimpleFunctions — what changed in the last hour
curl -s "https://simplefunctions.dev/api/changes"

API Comparison

Which platform fits your use case?

KalshiPolymarketMetaculusManifold
Auth for market dataNoNoNoNo
Auth for tradingAPI key + RSAWallet signatureN/AAPI key
WebSocketYesYesNoNo
Rate limit100/min500/minUndocumented1000/min
Historical dataSettlement CSVSubgraph + APIExportFull API
Real moneyYes (CFTC)Yes (USDC)NoSweepstakes
Market creationNo (Kalshi only)No (UMA only)CommunityAnyone
Best forUS regulated, macroCrypto-native, globalResearch, calibrationExperimentation

Contents


Exchanges & Platforms

Active prediction market platforms with programmatic access.

PlatformAPIMarketsSettlement
KalshiREST + WebSocketEconomics, politics, weather, scienceCFTC-regulated, USD
PolymarketCLOB APIPolitics, crypto, global eventsPolygon, USDC
MetaculusAPIScience, technology, geopoliticsReputation-based
Manifold MarketsAPIAnything (user-created)Play money + sweepstakes
PredictItAPIUS politicsCFTC no-action letter, USD
SmarketsAPIPolitics, sports, current affairsUK-regulated, GBP
FutuurAPIGlobal eventsPlay money + real money
Insight PredictionAPIGlobal eventsCrypto

APIs & SDKs

Multi-Venue

  • SimpleFunctions API — Unified REST API for Kalshi + Polymarket. Thesis management, edge detection, what-if scenarios, track record feedback. Free during beta.
  • SimpleFunctions CLI — 42 terminal commands for prediction market intelligence. Scan, watch, trade, agent mode.

Kalshi

Polymarket

Metaculus

Manifold

MCP Servers

Connect LLMs to prediction market data via the Model Context Protocol.

  • SimpleFunctions MCP — 25 tools: thesis management, edge detection, market search, trading, what-if scenarios. Works with Claude Code, Cursor, Windsurf.
    claude mcp add simplefunctions --url https://simplefunctions.dev/api/mcp/mcp
    
  • prediction-market-mcp-example — Minimal 50-line MCP server for learning. Search markets, get context, price changes.
  • us-gov-open-data-mcp — 300+ tools across 40+ U.S. government APIs. FRED, Treasury, BLS, Congress, FDA, EPA, SEC. Selective module loading.
  • fred-mcp-server — FRED economic data for Claude/Cursor.
  • mcp-fredapi — FRED API with series search and category browsing.
  • imf-data-mcp — IMF economic data via SDMX 3.0 API.

CLI Tools

  • SimpleFunctions CLI — sf scan, sf edges, sf watch, sf agent, sf intent. Covers Kalshi + Polymarket.
    npm i -g @spfunctions/cli && sf setup
    
  • polymarket-cli — Official Polymarket CLI for order placement.

Agent Frameworks & Examples

Building AI agents that reason over prediction market data.

Frameworks

Example Projects

Data & Datasets

Live Data

Historical Datasets

Aggregators & Cross-Venue

  • SimpleFunctions — Kalshi + Polymarket unified. Edge detection across venues.
  • MetaForecast — Aggregates Metaculus, Polymarket, Manifold, Good Judgment, and more.
  • Oddpool — Cross-venue odds, spreads, liquidity, arbitrage detection.
  • BaseRateTimes — News through the lens of prediction markets.
  • Electionbettingodds.com — Aggregated political prediction market odds.

Economic & Macro Data

The context layer for prediction markets — the economic indicators, government data, and macro signals that drive the events prediction markets trade on.

Aggregator Platforms

  • OpenBB — Open-source financial data platform (65K+ stars). Covers FRED, World Bank, BLS, Eurostat, and hundreds of other data sources. Python, CLI, and AI agent integration.
  • pandas-datareader — Extract data from FRED, World Bank, OECD, Eurostat, Stooq, and more into pandas DataFrames. The standard Python library for economic data access.
  • DBnomics — Aggregates 80+ statistical providers (central banks, national statistics offices, international organizations). R client, Julia client, Stata client. Free API, no auth.
  • findatapy — Download market data from Bloomberg, Eikon, Quandl, FRED and more via a unified interface.

FRED (Federal Reserve Economic Data)

  • fredapi — Python client for FRED. pip install fredapi. Supports series search, vintages, releases.
  • fredr — R client for FRED API. On CRAN.
  • fred-rs — Rust crate for FRED.
  • fredcpp — C++ client for FRED.
  • FredApi.jl — Julia package for FRED.
  • fredric — Ruby library and CLI for FRED.
  • fred-mcp-server — MCP server for FRED data in Claude/Cursor.
  • mcp-fredapi — Another FRED MCP server with search and category browsing.

Federal Reserve & Central Banks

  • FRB — Python client for all Federal Reserve Bank data (not just FRED — includes flow of funds, financial accounts, survey data).
  • econ_data — Python notebooks demonstrating FRED, BLS, Census, and BEA API usage with working examples.

World Bank

  • wbdata — Python library for World Bank data. Tested: pulls GDP per capita for any country/date range. No auth.
  • world_bank_data — Python library with pandas integration. Tested: pulls CPI inflation for US/CN/GB/BR. No auth.
  • WDI — R package for World Bank data. On CRAN, 250+ stars.
  • WorldBankData.jl — Julia package for World Bank data.

BLS (Bureau of Labor Statistics)

  • blscrapeR — R package for BLS data (unemployment, CPI, wages). 117 stars.
  • blsAPI — R client for BLS API. 93 stars.
  • BLS-APIs — Python wrapper for BLS v1/v2 APIs.
  • BLS API v1 requires no auth key: curl https://api.bls.gov/publicAPI/v1/timeseries/data/LNS14000000

IMF

  • IMFData — R package for IMF data. 49 stars.
  • imf-data-mcp — MCP server for IMF economic data via SDMX 3.0 API.
  • IMF API requires no auth: https://data.imf.org (SDMX 3.0)

Eurostat

  • eurostat-api-client — Python client for Eurostat data.
  • eurostat_downloader — Python tool for bulk Eurostat downloads.
  • Eurostat API requires no auth: https://ec.europa.eu/eurostat/api/dissemination/statistics/1.0/data/

Multi-Agency / Government

  • us-gov-open-data-mcp — MCP server + TypeScript SDK for 40+ U.S. government APIs (300+ tools). Covers FRED, Treasury, BLS, Census, EPA, FDA, SEC, Congress, and more. 20+ APIs need no auth key. 94 stars.

OECD

  • OECD Data API requires no auth: https://sdmx.oecd.org/public/rest/data/ (SDMX format)

Analytics & Research Tools

Tutorials & Guides

Getting Started

Building Agents

Trading Systems

Orderbooks & Market Microstructure

Edge Detection & Sizing

Strategy & Market Structure

How Prediction Markets Work

Concepts & Frameworks

Platform Comparisons

Selected Reading

Primers

Strategy & Structure

AI & Prediction Markets

Academic Papers

Project Ideas

Things you can build with the APIs and tools above.

ProjectDifficultyStack
Cross-venue arbitrage detector — find price differences between Kalshi and Polymarket on the same eventMediumPython + both APIs
Telegram alert bot — notify on 10%+ price movesEasyNode.js + SimpleFunctions /api/changes
Thesis-driven trading agent — create a causal model, scan for edges, auto-tradeHardLangChain + SimpleFunctions MCP
Prediction market dashboard — real-time prices, orderbook depth, historical chartsMediumReact + WebSocket APIs
Forecasting tournament bot — auto-submit forecasts to Metaculus/Manifold using LLMsMediumPython + OpenAI + Metaculus API
Market sentiment tracker — correlate prediction market moves with news/X sentimentHardPython + SimpleFunctions + X API

Community


Contributing

  1. Fork this repo
  2. Add your resource to the appropriate section
  3. Keep descriptions short (one line)
  4. Include a link to source code or documentation
  5. Submit a PR

What belongs here: APIs, SDKs, datasets, code, tools, research, papers, tutorials. Things developers use to build.

What doesn't belong: Consumer dashboards, portfolio trackers, copy-trading services, news aggregators without an API. Check Awesome Prediction Market Tools for those.

License

MIT

Ver o original no GitHub

Bom para

Relacionados