convez.ai
MCP ServerFinanceiroAvançado
Alisammour

Storyflo MCP — Camada de Sinal para Trading Algorítmico

Alisammour/storyflo-algo-trader-mcp-examples

Sinal jornalístico estruturado para algo traders em mercados de previsão

Ver no GitHub
InicialPluginScore · de 0 a 100

MCP server que fornece contexto jornalístico estruturado para trading bots em Polymarket, Kalshi e IBKR. Integra briefings diários, busca de artigos vinculados a contratos e análise de sentimento counter-consensus para alimentar decisões de entrada com fundamentação editorial verificada.

Instalação

$npx convez-ai Alisammour/storyflo-algo-trader-mcp-examples

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
0
Acompanhando
0
Saúde
42%
Último commit
há 5 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Gate de entrada em operações Polymarket/Kalshi usando counter-consensus scoring
  • Análise de briefings macro antes de trades de eventos econômicos (CPI, Fed)
  • Geração de relatório diário de alpha vinculando notícias a contratos ativos

Veredito

Solução robusta para traders que precisam contextualizar movimentos de mercado com narrativa editorial. A arquitetura MCP é clara e os 3 exemplos (Polymarket signal bot, Kalshi macro gate, daily alpha feed) são implementações production-ready. Restrição de compliance e modelo de monetização x402 tornam a integração legítima e escalável.

README

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

storyflo-mcp signal layer for Polymarket / Kalshi / IBKR ForecastTrader bots

TL;DR: storyflo-mcp gives algo traders the cited-journalism signal layer Polymarket and Kalshi don't ship. Three working examples show how to use it in your trading loop.

If you're building a Polymarket / Kalshi / IBKR ForecastTrader bot, you've got the raw event-contract odds. What you don't have is a structured way to answer "why is this market moving right now, and which articles is the consensus reading?"

Storyflo runs four named hosts producing a daily audio + text brief across tech / markets / news / sports / health verticals. The free MCP server at api.storyflo.com/v1/mcp/sse exposes that as 8 tool calls — including get_market_linked_stories, get_vertical_briefing, search_articles, digest. Storyflo also matches every story we publish to a live Kalshi or Polymarket event contract via the market_context worker, so the signal layer comes pre-aligned with the markets you're trading.

Counsel-cleared posture: storyflo cites event-contract probabilities as journalist input, never as buy-now output. Your bot stays compliant with each platform's terms; we never redistribute odds as display.


Quick start

# 1) Install the storyflo MCP client
npm i -g storyflo-mcp@latest

# 2) Get the examples
git clone https://github.com/Alisammour/storyflo-algo-trader-mcp-examples.git
cd storyflo-algo-trader-mcp-examples

# 3) Run any of the 3 examples
python3 polymarket_signal_bot.py
python3 kalshi_macro_bot.py
python3 daily_alpha_feed.py

No STORYFLO_TOKEN required for the tools/list discovery calls. tools/call works anonymously for the free tools (search_articles, get_vertical_briefing, digest, get_market_linked_stories, get_audio_url); paid tools (get_premium_briefing) settle in USDC on Base via x402.


What's in the box

FileWhat it shows
polymarket_signal_bot.pyPoll a Polymarket market by slug → call get_market_linked_stories → score the sentiment + flag counter-consensus signals. Use the flag as a gate on your existing entry logic.
kalshi_macro_bot.pyOn a CPI / Fed / jobs / GDP print, call get_vertical_briefing(vertical="finance") → Mason's macro brief → use the brief's signal-tags (high_velocity, genuine_uncertainty) as a trade gate against your Kalshi positions.
daily_alpha_feed.pyCron-style daily — call digest(verticals=["finance","crypto"]) → return a markdown alpha report listing every story that matches a Kalshi or Polymarket contract from the last 24h.

The 8 storyflo-mcp tools

ToolReturnsWhen to call from a trading loop
search_articles(q, vertical?, limit?)Recent articles matching queryBuild a "macro reads" feed for your model
get_article(slug)Single story with body + audio_urlPull the full context for a flagged trade
get_audio_url(slug)R2 audio URLGenerate a daily audio brief for your team
get_vertical_briefing(vertical, lang?)Mason / Riley / Theo / Iris's daily briefMacro / sports / tech / health pre-trade read
get_market_linked_stories(slug)Stories paired to a market contractReal-time why is this moving lookup
subscribe_topic(topic)Subscribe a topic for the digestBuild a watch-list that powers digest()
list_subscriptions()List your subscribed topicsInspect your watch-list state
digest(verticals?)Daily digest of subscribed topicsCron-style daily roll-up

Why storyflo and not just scraping headlines

  • Structured, not scraped. Each tool returns typed JSON. No HTML.
  • Counsel-cleared posture. We cite markets-as-INPUT (journalism), never markets-as-OUTPUT (bet-now). Your bot stays compliant with Polymarket / Kalshi terms.
  • Pre-aligned with the markets you trade. Our market_context worker already matches every story we publish to a live Kalshi / Polymarket contract.
  • Persona attribution. Mason on markets ($Fed, $CPI, $jobs); Theo on tech; Riley on sports + politics + world; Iris on health + longevity + science.
  • Free. The 7 listed tools work anonymously or via OAuth. Premium briefings settle in USDC on Base via x402 — get_premium_briefing.
  • MCP-native. Drop straight into Claude / ChatGPT / Cursor / VS Code via MCP — or call the HTTP/SSE endpoint at api.storyflo.com/v1/mcp/sse directly.

Useful links

  • Storyflo home: storyflo.com
  • MCP endpoint: api.storyflo.com/v1/mcp/sse
  • npm package: storyflo-mcp
  • Daily personas: storyflo.com/personas — Theo on tech · Mason on markets · Riley on sports + news · Iris on health
  • Issues / requests: open a GitHub issue here or email ops@storyflo.com

License

MIT — see LICENSE. Copy any of these examples directly into your trading code; no attribution needed beyond the storyflo MCP endpoint URL.

— Best, Ali, storyflo.com

Ver o original no GitHub

Bom para

Relacionados