Skip to main content
Glama
tiezhengzhang

hk-ipo-fee-mcp

hk-ipo-fee-mcp

PyPI License: MIT

An MCP server that gives AI assistants exact Hong Kong IPO cost maths, so they can stop guessing. It covers:

  • the entry fee (amount payable), including the 1% brokerage, SFC, AFRC and trading-fee levies, each rounded to the cent;

  • margin-financing interest, the break-even sell price, scenario P&L and expected value;

  • selling costs, with stamp duty rounded up to a whole HK$;

  • cited rule summaries (Mechanism A/B allocation, Pool A/B, FINI T+2);

  • open-licensed (Wikidata, CC0) lookups of HK-listed companies and listing dates.

中文简介:港股打新费用计算 MCP 服务器。让 Claude、Cursor 等 AI 助手精确计算入场费(1% 佣金 + 证监会/会财局征费 + 交易费,逐项四舍五入至分)、孖展利息、回本价、情景盈亏与期望值、卖出费用(印花税向上取整至元),并查询机制 A/B、甲乙组、FINI 等规则(附出处)。公司与上市日期查询只使用 Wikidata(CC0)开放数据,不抓取港交所/披露易。免费开源(MIT)。网页版计算器:https://hk-ipo-calc.pages.dev/

Ask things like:

"I'm applying for 20 lots of an IPO at HK$8.88, 500 shares a lot, with 10× margin at 3.5% for 7 days and a HK$100 handling fee. 18,000 out of 120,000 applicants at my size got one lot. What's my entry fee, break-even price and expected value?"

The numbers come from a tested fee engine, not from the model. That engine is the same one used by the web calculator and the HK IPO Ledger Chrome extension, and the two implementations are checked to give identical results on 8,000 test values.

Tools

Tool

What it does

hk_ipo_amount_payable

Amount payable for N shares: application money + 1% brokerage + 0.0027% SFC levy + 0.00015% AFRC levy + 0.00565% trading fee

hk_ipo_plan

Full plan: entry fee, total payable, own cash vs margin loan, interest, fixed cost, break-even price, scenario table, expected value (from allotment counts or a rate)

hk_sell_costs

Selling costs: 0.1% stamp duty rounded up to HK$1, trading fee, levies, broker commission / platform fee

hk_ipo_rules

Cited rule summaries: fees, allocation (Mechanism A/B from 4 Aug 2025, 18C), pools, timeline (FINI) or all

hk_company_lookup

Look up an HK-listed company by code or name (Wikidata, CC0)

hk_listings

Listings by date range (Wikidata, CC0, cached 24 h)

Related MCP server: hk-markets-mcp

Install

You need Python ≥ 3.10. The easiest way to run it is with uv, which needs no install step:

uvx hk-ipo-fee-mcp          # runs the stdio server
# or
pipx install hk-ipo-fee-mcp && hk-ipo-fee-mcp

Claude Desktop

Edit claude_desktop_config.json. On macOS it is at ~/Library/Application Support/Claude/; on Windows at %APPDATA%\Claude\.

{
  "mcpServers": {
    "hk-ipo-fee": { "command": "uvx", "args": ["hk-ipo-fee-mcp"] }
  }
}

Claude Code

claude mcp add hk-ipo-fee -- uvx hk-ipo-fee-mcp

Cursor

Add this to ~/.cursor/mcp.json, or to .cursor/mcp.json in a project:

{
  "mcpServers": {
    "hk-ipo-fee": { "command": "uvx", "args": ["hk-ipo-fee-mcp"] }
  }
}

VS Code (GitHub Copilot agent mode)

Add this to .vscode/mcp.json:

{
  "servers": {
    "hk-ipo-fee": { "type": "stdio", "command": "uvx", "args": ["hk-ipo-fee-mcp"] }
  }
}

Windsurf / Cline / other clients

Any client that speaks MCP over stdio works. Set the command to uvx and the args to ["hk-ipo-fee-mcp"]. If you installed with pipx, set the command to hk-ipo-fee-mcp with no args.

Environment variables (optional)

Variable

Default

Purpose

HKIPO_CONTACT

project URL + maintainer email

Contact string in the Wikidata User-Agent (Wikimedia UA policy). Set your own if you query a lot

HK_IPO_FEE_MCP_CACHE

~/.cache/hk-ipo-fee-mcp

Cache directory for listing data

Example

hk_ipo_plan(offer_price=8.88, board_lot=500, lots_applied=20, margin_multiple=10, annual_rate=0.035, interest_days=7, handling_fee=100, allotted_lots=1, valid_applications=120000, successful_applicants=18000) returns:

Entry fee (1 lot)

HK$4,484.78

Total amount payable (20 lots)

HK$89,695.55

Margin interest

HK$54.19

Fixed cost (paid either way)

HK$154.19

Break-even sell price

HK$9.289 (+4.61%), assuming zero sell commission

(These are illustrative inputs, not a real IPO.)

Data & compliance

  • Fees and rules follow published SFC, AFRC, HKEX and IRD (GovHK) rates; every rule summary cites its source.

  • No HKEX/HKEXnews scraping. HKEX's terms prohibit automated access, so IPO terms (offer price, board lot, allotment results) must come from you or the prospectus. The server never invents them.

  • Listing lookups use Wikidata (CC0). The data is community-maintained, so verify it before relying on it.

  • Results are arithmetic, not investment advice. This project is not affiliated with Hong Kong Exchanges and Clearing Limited.

Development

git clone https://github.com/tiezhengzhang/hk-ipo-fee-mcp && cd hk-ipo-fee-mcp
python -m venv .venv && . .venv/bin/activate && pip install -e . pytest
python -m pytest -q                 # fee anchors + MCP stdio end-to-end

More from the author

  • 🧮 HK IPO Calculator — a free bilingual web calculator using the same fee engine, plus guides on Mechanism A/B, Pool A/B, margin and fees.

  • 📒 HK IPO Ledger (Chrome extension) — an IPO ledger across multiple brokers, with allotment and P&L stats, kept on your device. Free for up to 30 records. Pro is a one-time payment with no subscription.

License

MIT © 2026 Tiezheng Zhang. Listing data from Wikidata is CC0.

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    B
    maintenance
    Enables AI assistants to fetch realtime stock quotes, fundamentals, and search for stocks across A-share, Hong Kong, and US markets without requiring API keys.
    5
    5 npm
    MIT
  • A
    license
    A
    quality
    C
    maintenance
    Enables AI agents to retrieve live Hong Kong monetary data (HKMA Base Rate, HIBOR fixings, USD/HKD peg band) via typed tools, no API key needed.
    3
    7 npm
    MIT
  • A
    license
    Not graded
    quality
    A
    maintenance
    Enables AI assistants to retrieve live Hong Kong banking data—ATM and branch locations, HIBOR/HKD rates, and bank legitimacy checks—from official HKMA public APIs, with no credentials or scraping.
    MIT
  • A
    license
    Not graded
    quality
    A
    maintenance
    Enables AI assistants to fetch real-time quotes, K-line data, and technical indicators for A-share, Hong Kong, US stocks, funds, futures, and options, plus composite stock analysis, market scanning, and sector breakdowns. It also ships preset prompts and skills for technical analysis, stock screening, market overviews, and portfolio monitoring.
    109 npm
    ISC