AI导哈

机构大宗交易数据Python客户端与MCP服务器

非官方的Python客户端和MCP服务器，用于从volumeleaders.com获取机构大宗交易、交易聚类、耗竭得分等数据，支持量化分析和AI编码助手等场景。

  • by major
  • Python
  • 本地部署
  • 3 星标
  • 更新于
在 GitHub 查看 major/volumeleaders

使用说明

Python client and MCP server for volumeleaders.com

Unofficial. This project is not affiliated with, endorsed by, or connected to volumeleaders.com in any way.

Python client and MCP server for volumeleaders.com institutional block trade data.

Use it as a Python library to query institutional block trades, trade clusters, trade levels, and more with typed Pydantic models. Or run it as an MCP server to give AI coding assistants direct access to the same data.

Authentication works by extracting cookies from your browser session (no API keys needed).

Prerequisites

You must be logged into volumeleaders.com in Firefox before using the client. The library reads your browser cookies for authentication (there is no programmatic login).

Install

pip install volumeleaders

Or with uv:

uv add volumeleaders

Usage

from volumeleaders import VolumeLeadersClient, get_trades, get_exhaustion_scores

with VolumeLeadersClient() as client:
    # Fetch institutional trades for a date range
    trades = get_trades(
        client,
        start_date="2026-03-01",
        end_date="2026-03-31",
        tickers="AAPL",
    )
    for trade in trades:
        print(f"{trade.ticker} {trade.trade} @ {trade.current}")

    # Get exhaustion scores for today
    scores = get_exhaustion_scores(client)
    print(scores)

Available endpoints

Function Description
get_trades Institutional block trades with full filter set
get_trade_clusters Clustered trade activity
get_trade_cluster_bombs Large cluster events
get_exhaustion_scores Daily exhaustion score ranks
get_earnings Earnings data
get_company Company metadata
get_snapshot Chart snapshot for a ticker
get_all_snapshots Price snapshots for all tickers
get_price_data Price/volume bar data
get_chart_trades Individual large trades from the chart envelope
get_chart_levels Chart trade levels
get_trade_levels Trade level data
get_trade_level_touches Trade level touch events
get_institutional_volume Institutional volume
get_ah_institutional_volume After-hours institutional volume
get_total_volume Total volume
get_trade_alerts Trade alerts
get_trade_cluster_alerts Trade cluster alerts
get_alert_configs Alert configurations
get_watchlist_configs Watchlist configurations
get_watchlist_tickers Tickers in a watchlist
get_sector_breakdown Daily institutional dollar volume by sector
get_institutional_outliers Statistical block trade volume outliers
get_notional_by_sector_by_name Hierarchical sector, theme, and ticker capital flows
get_supply_demand_areas Automated sector supply and demand support scores
get_sector_daily_returns Multi-factor sector metrics, momentum, Sharpe, and risk
get_dark_pool_volume_report Binned dark pool volume distribution profiles

All endpoint functions take a VolumeLeadersClient as the first argument and return typed Pydantic models.

MCP Server

The library includes an MCP server that exposes VolumeLeaders data as tools for AI coding assistants. 12 tools are available:

  • trades: Institutional block trade scanner with context-aware defaults.
  • trade_clusters: Institutional trade cluster scanner for a given day.
  • trade_cluster_bombs: Large trade cluster bomb events over date ranges.
  • trade_levels: Institutional trade levels for a ticker symbol.
  • trade_level_touches: Intraday trade level touch scanner.
  • sector_flows: Daily institutional dollar flows and market share across sectors.
  • institutional_outliers: Statistical block trade volume anomalies (Z-scores).
  • sector_themes: Hierarchical capital allocation (Sector -> Theme -> Tickers).
  • sector_support_scores: Automated supply and demand support score distributions.
  • sector_factors: Multi-factor sector scorecard (Momentum, Sharpe, Beta, Volatility).
  • sector_rotation_rrg: JdK Relative Rotation Graph (RRG) modeling against SPY.
  • dark_pool_profile: Dark pool price-volume profile and Point-of-Control (POC).

Quick start with uv

No installation needed:

uv run --with "volumeleaders" volumeleaders-mcp

Or if the package is already installed:

volumeleaders-mcp

Client configuration

Claude Code (.mcp.json)

Create .mcp.json in your project root.

Using uv:

{
  "mcpServers": {
    "volumeleaders": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "--with", "volumeleaders", "volumeleaders-mcp"]
    }
  }
}

After adding the config, restart Claude Code and run /mcp to verify the server appears.

OpenCode (opencode.json)

Add to opencode.json (or opencode.jsonc) in your project root.

Using uv:

{
  "mcp": {
    "volumeleaders": {
      "type": "local",
      "command": ["uv", "run", "--with", "volumeleaders", "volumeleaders-mcp"]
    }
  }
}

License

MIT

相关 MCP