Skip to main content

MCP data layer for AI-assisted market research: 10 finance tools for AI agents covering DeFi yield, Hyperliquid vaults, SEC Form 4 insider clusters, 8-K material filings, macro tape, US spot ETF flows, airdrop farms, Polymarket whales, confluence_today cross-source signal alignment, and onchain_smart_wallets (live Solana copy-trade bot score data). Free tier requires no signup. Pro $19/mo for real-time.

Project description

falsifylab-alpha-mcp

smithery badge PyPI License

MCP data layer for AI-assisted market research. 10 live finance tools that drop into Claude Code, Cursor, Cline, Windsurf, or any MCP-compatible client. Free tier requires no signup.

Not a signal service. Not investment advice. Public market data with light enrichment for AI agents. Backtest before acting.

Two ways to use it

Option 1: hosted MCP (zero install). Point your agent at the hosted endpoint, no Python needed.

{
  "mcpServers": {
    "falsifylab-alpha": {
      "url": "https://mcp.falsifylab.com/"
    }
  }
}

Free tier auto-applies. OAuth 2.1 + PKCE flow for paid plan via Whop. Streamable HTTP transport. Full OpenAPI 3.1 spec at https://docs.falsifylab.com/openapi.html.

Option 2: stdio (this package).

pip install falsifylab-alpha-mcp

Zero runtime dependencies (Python stdlib only). Then wire to your agent. see client setup.

What it does

When the user asks your AI agent things like:

  • "What insider clusters are forming today?"
  • "Show me Hyperliquid vaults with sharp Sharpe and <10% drawdown"
  • "Are BTC ETF flows confirming or diverging from price action?"
  • "What's in the macro tape right now?"
  • "Find Polymarket whales with 60%+ win rate"

...the agent calls one of 10 tools and grounds its answer in real-time numbers from public-market data — instead of speculating from training-data knowledge.

The 10 tools

Tool Returns Use case
top_yield_farms DeFi yields with emissions stripped, IL risk priced Find real yield, not headline yield
hl_vault_leaderboard Hyperliquid vaults by NAV / 30d / drawdown / score Copy-trade reference
insider_buy_clusters SEC Form 4 cluster buys (3+ insiders, 24h window) Equity conviction signal
sec8k_material_today 8-K filings filtered by material item codes Earnings, M&A, officer changes
macro_tape SPX/NDX/VIX/UST/DXY/GOLD/WTI/BTC/ETH snapshot Regime gate for trade ideas
etf_flow_today US spot BTC + ETH ETF aggregates Institutional positioning
active_airdrop_farms DefiLlama yield-gap airdrops Farms where you're paid to wait
polymarket_whale_positions Top wallets + active positions Prediction-market copy-trade
confluence_today Cross-source signal alignment Stacks 2+ signals on same asset
onchain_smart_wallets Live Solana copy-trade bot scores Production wallet rankings

All tools have full JSON schemas exposed at /.well-known/mcp/server-card.json.

Client setup

Claude Code

claude mcp add falsifylab-alpha -- python -m falsifylab_alpha_mcp

For Pro tier:

claude mcp add falsifylab-alpha \
  --env FL_API_KEY=fl_your_key \
  -- python -m falsifylab_alpha_mcp

Cursor

Drop into ~/.cursor/mcp.json or workspace root .mcp.json:

{
  "mcpServers": {
    "falsifylab-alpha": {
      "command": "python",
      "args": ["-m", "falsifylab_alpha_mcp"],
      "env": {
        "FL_API_KEY": "fl_your_key_or_omit_for_free_tier"
      }
    }
  }
}

Restart Cursor. Tools auto-discovered.

Cline

Settings → MCP Servers → Add Server. Paste the same JSON as Cursor above.

Windsurf

Settings → MCP. Same JSON config. Auto-indexed from Smithery within 48h.

Smithery

Connect via Smithery's hosted gateway — no local Python install needed. See smithery.ai/servers/falsifylab/falsifylab-alpha-mcp.

Verify install

In your agent, ask:

Use macro_tape. Show me SPX, VIX, BTC, ETH with 1d and 5d returns.

If the agent returns numbers, you're live.

Working agent prompts

Copy-paste these. They work on first run.

Daily macro brief:

Use macro_tape. Tell me the macro regime in 3 sentences. Cite specific
numbers — last price, 1d return, 5d return. Then list 2 tickers from
sec8k_material_today that might benefit from this regime.

Insider + 8-K stacker:

Run insider_buy_clusters (min_insiders=3) and sec8k_material_today
(items=["2.02","8.01","5.02"]). Find tickers in both within 24h. For
each match, summarize what insiders bought and what the 8-K said. End
with the bear case in one sentence.

Confluence scan with bear case:

Use confluence_today with min_signals=2. For each asset/ticker, explain
why these signals stacking is meaningful in 2 sentences. Skip noise
(DeFi LP pools with stablecoins). Highlight the most contrarian stack.

8 more recipes at falsifylab.com/cookbook.

Pricing

Tier Price What
Free $0 24h cached, 10 results/query, no signup, 60 req/hr
Pro $19/mo Real-time (5min), 100 results/query, 90d history, no rate limit
Pro Plus $49/mo 1-min refresh, email + Slack alerts, 365d on equity feeds, webhooks
Teams $199/mo 5 seats, 50k req/day, priority support

Launch promo EARLY50: 50% off first 3 months, capped 25 redemptions. Get a key at falsifylab.com/pro.

Live demo

Public agent calling these tools every 15 minutes: falsifylab.com/demo. Same MCP server you'd install. Real numbers, no curation.

Why this exists

Built it because the author hand-scraped SEC EDGAR and DefiLlama at 6am every morning to feed a fleet of trading bots. The bots that survived ate from this exact pipeline. Wrapped it as an MCP server so any agent can pull the same data.

The bot fleet receipts (alive + dead, P&L, drawdown) are public at falsifylab.com/vaults.

Transparency

  • What we do: aggregate + lightly enrich public market data, expose via MCP
  • What we don't: trade execution, brokerage, advisory, signal selling
  • Data sources: SEC EDGAR (Form 4, 8-K), SoSoValue (ETF flows), DefiLlama (yield), Hyperliquid public API, Polymarket public API, Yahoo Finance (macro)
  • Privacy: free-tier key issued by hash(IP+UA), not stored long-term. Request bodies never logged.

Architecture

your AI agent
    ↓ (MCP stdio)
falsifylab_alpha_mcp (Python stdlib, no deps)
    ↓ (HTTPS)
falsifylab.com/api/* (Cloudflare Workers)
    ↓ (cached 60s-1hr)
R2 (data feeds refreshed every 15min via cron)
    ↓ (every 15min)
Backend cron scripts (Form 4, 8-K, ETF, yield, polymarket, onchain, etc.)

Free tier hits cached data with 24h age. Pro tier hits the same endpoints with auth header, bypasses cache to fresh upstream.

Links

Support

Contributing

PRs welcome for:

  • New tool ideas (open issue first to discuss scope)
  • Documentation improvements
  • Bug fixes

Do not submit PRs for trading-signal generators or copy-trading helpers — out of scope for this server (we are research data infrastructure, not advisory).

License

MIT. See LICENSE.


FalsifyLab Alpha v0.1.5 · Built for AI coding agents · Free tier no signup · falsifylab.com

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

falsifylab_alpha_mcp-0.2.2.tar.gz (10.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

falsifylab_alpha_mcp-0.2.2-py3-none-any.whl (10.1 kB view details)

Uploaded Python 3

File details

Details for the file falsifylab_alpha_mcp-0.2.2.tar.gz.

File metadata

  • Download URL: falsifylab_alpha_mcp-0.2.2.tar.gz
  • Upload date:
  • Size: 10.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.3

File hashes

Hashes for falsifylab_alpha_mcp-0.2.2.tar.gz
Algorithm Hash digest
SHA256 813a5af109225efb41f0f9d85c75c84e70589198cd7f1f78f04b747b3e4f5dbc
MD5 e4810c679a3b96ec8a72cc2956465d6f
BLAKE2b-256 9019cdd729a8ac9c05d95ec630831ba752675653d6322db0957d59979033a370

See more details on using hashes here.

File details

Details for the file falsifylab_alpha_mcp-0.2.2-py3-none-any.whl.

File metadata

File hashes

Hashes for falsifylab_alpha_mcp-0.2.2-py3-none-any.whl
Algorithm Hash digest
SHA256 a08fbfc1f1c347d8a7ab07078a485b49ee0ff68a89fc3519443f59b8ee13b557
MD5 0d0c7bbd86e41e8c58d4afe0b7cc3050
BLAKE2b-256 7339210dad5cf21afc8390033f48f318a9d18761ea33508f514528ef8fba73d8

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page