Tapeline MCP Server

Tapeline

A live quantitative market scanner for retail traders.

Tapeline surfaces high-conviction technical and fundamental signals across US stocks and ETFs. It scores ~2,500 tickers every minute during market hours, detects Bollinger Band squeezes and volume expansions, tracks Congressional trades and institutional flows, and classifies the overall market regime.

Built on the same engine that powers a production personal trading bot.

---

Status

Pre-launch scaffold. This repo is a fresh clone separate from the personal C:\signal-system\ engine. Nothing here is wired to live data yet; that's the first build milestone.

Architecture (see docs/ARCHITECTURE.md)

┌──────────────────────┐      ┌──────────────────┐      ┌─────────────────┐
│  Massive SIP feed    │─────▶│  Scoring worker  │─────▶│  Postgres       │
│  (Polygon, rebranded │      │  (adapted from   │      │  (scores,       │
│   2025-10-30)        │      │   signal-system) │      │   snapshots)    │
└──────────────────────┘      └──────────────────┘      └────────┬────────┘
                                                                 │
                                                                 ▼
┌──────────────────────┐      ┌──────────────────┐      ┌─────────────────┐
│  Next.js dashboard   │◀─SSE─│  FastAPI         │◀─────│  Read API       │
│  (scanner, squeeze,  │      │  (auth, billing, │      │                 │
│   regime, congress)  │      │   live stream)   │      │                 │
└──────────────────────┘      └──────────────────┘      └─────────────────┘
          │                            │
          ▼                            ▼
┌──────────────────────┐      ┌──────────────────┐
│  Clerk (auth)        │      │  Stripe (billing)│
└──────────────────────┘      └──────────────────┘

Stack

• Backend: Python 3.12 + FastAPI + SQLAlchemy + Alembic
• Database: Postgres (Supabase or Neon)
• Frontend: Next.js 14 + TypeScript + Tailwind + shadcn/ui
• Auth: Clerk
• Billing: Stripe
• Data feed: Massive Stocks Starter ($29/mo, commercial redistribution) — formerly Polygon.io
• Macro / fundamentals / 13F: FRED · Finnhub · Quiver Quantitative
• Deployment: Fly.io (backend + worker) + Vercel (frontend)
• Email: Resend

Product tabs (v1)

1. 📡 Scanner — ~2,500 tickers, composite score, filters, sort
2. 🔥 Squeeze Watch — BB squeeze days, volume expansion, OBV trend, suggested window
3. 🌊 Market Regime — current regime, VIX, DXY, 10Y, rate direction, sector leaders
4. 🏛️ Congress Trades — recent politician buys/sells aggregated by ticker

Pricing (planned)

• Free: daily delayed snapshot, 10-ticker scanner
• Pro ($29/mo): live updates, full scanner, squeeze + regime, email alerts
• Premium ($49/mo): + Congress, Telegram alerts, CSV export, API access

Repo layout

backend/          FastAPI + scoring worker
frontend/        Next.js dashboard (initialize with `npx create-next-app@latest`)
infra/           Docker, deployment config
docs/            Architecture, legal checklist, data sources, pricing
scripts/         One-off ops scripts

Getting started

See docs/ARCHITECTURE.md for the full technical plan and docs/LEGAL_CHECKLIST.md for pre-launch must-dos.

What this is NOT

Tapeline is a quantitative scanning and research tool. It does not:

• Provide individualized investment advice
• Execute trades
• Make price predictions
• Manage customer funds

All output is factual data synthesis. Users make their own decisions. See docs/LEGAL_CHECKLIST.md.