Webull MCP Server
Enables MCP clients to read Webull account balances, positions, orders, stock quotes, historical bars, and US option market data, with live equity trading available only as an explicitly enabled opt-in.
README
Webull MCP Server
An MCP server that exposes the official
Webull OpenAPI (webull-openapi-python-sdk)
as tools an MCP client — Claude Code, Claude Desktop, or any other — can call.
Read account balances, positions, orders, stock quotes, historical bars, and
US option market data; optionally place and cancel live equity orders.
⚠️ Read this before you install
Unofficial. This project is not affiliated with, endorsed by, or supported by Webull. It calls Webull's OpenAPI as a third-party client, at your own risk. The API can change or break without notice, and nothing here is guaranteed to keep working.
Not financial advice. This is plumbing, not a strategy. It has no opinion about what you should buy or sell. You are solely responsible for every order that leaves your account and every dollar it costs you.
Trading is disabled by default — and there is no paper mode. The order
tools are not even registered with the MCP client unless you explicitly set
WEBULL_ENABLE_TRADING=true. Webull publishes no paper-trading endpoint,
so once enabled, every order this server places is live, with real money, on
a real account. See Enabling live trading.
Your credentials stay on your machine. Your App Key and Secret live only in
a local .env file, which is gitignored. They are sent only to Webull's API —
never to the model, never to any third party.
Tools
Read-only (always available)
| Tool | What it does |
|---|---|
list_accounts |
List accounts available to your credentials |
get_account_balance |
Cash, net liquidation, buying power |
get_positions |
Current holdings |
get_open_orders |
Working orders |
get_order_history |
Filled, cancelled, and past orders |
get_instrument |
Resolve a ticker to instrument metadata |
get_quote |
Real-time price snapshot (last, OHLC, volume) |
get_history_bars |
Historical OHLCV candles |
get_option_quote |
Snapshots for US option contracts — bid/ask, volume, open interest, IV, and all five Greeks |
get_option_history_bars |
Historical OHLCV bars for option contracts |
get_option_tick |
Tick-by-tick trades for a single option contract |
trading_status |
Reports whether live trading is enabled |
Option tools take OCC-format symbols (AAPL260619C00285000 —
<ROOT><YYMMDD><C|P><strike × 1000, 8 digits>). Webull has no chain-listing
endpoint, so build candidate symbols from known strikes and expirations and
quote them directly. Up to 20 per call.
Live trading (gated, off by default)
Registered only when WEBULL_ENABLE_TRADING=true:
| Tool | What it does |
|---|---|
preview_order |
Validate an order without submitting — fees, buying-power impact, validation errors |
place_order |
Submit a live equity order. Additionally requires confirm=True |
cancel_order |
Cancel a working order by client_order_id |
Supported order types: MARKET, LIMIT, STOP_LOSS, STOP_LOSS_LIMIT;
DAY or GTC; optional extended-hours routing.
Example session
What using it actually looks like from an MCP client. Values below are illustrative — substitute your own account.
> How's my account doing, and what's the SMCI Jan 250 call worth?
● get_account_balance()
⎿ {
"account_id": "XXXXXXXXXXXXXXXXXXXXXXXXXX",
"net_liquidation": "18432.55",
"cash_balance": "6120.18",
"buying_power": "12240.36"
}
● get_positions()
⎿ [
{ "symbol": "INTC", "quantity": "40", "cost_price": "127.54",
"last_price": "131.02", "unrealized_pnl": "139.20" },
{ "symbol": "IWM", "quantity": "15", "cost_price": "241.88",
"last_price": "238.15", "unrealized_pnl": "-55.95" }
]
● get_option_quote(symbols: "SMCI260116C00250000")
⎿ [
{ "symbol": "SMCI260116C00250000",
"bid": "12.40", "ask": "12.85", "volume": "1843",
"open_interest": "9021", "imp_vol": "0.5412",
"delta": "0.4187", "gamma": "0.0092", "theta": "-0.1734",
"vega": "0.2815", "rho": "0.0946" }
]
Net liq is $18,432.55 with $12,240 buying power. INTC is up $139 and IWM
is down $56. The SMCI 250 call is $12.40 × $12.85 — 0.42 delta, 54% IV,
and theta is costing you $17/day per contract.
Only read-only tools appear above.
place_orderis not available at all unless you explicitly enable trading — see Enabling live trading.
Setup
1. Get Webull API credentials
- Go to the Webull Developer portal and sign in with the Webull account you want the server to access.
- Apply for OpenAPI access and create an app. Approval is not instant — Webull reviews requests.
- Copy the App Key and App Secret it issues.
Your account must have OpenAPI enabled for your region. Real-time market data may additionally require a market-data entitlement on your account; without one, quotes may be delayed.
2. Install
git clone https://github.com/<your-username>/webull-mcp.git
cd webull-mcp
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
Requires Python 3.10 or newer.
3. Configure
cp .env.example .env
Edit .env and fill in your credentials:
WEBULL_APP_KEY=your_app_key_here
WEBULL_APP_SECRET=your_app_secret_here
# One of: us, hk, jp, sg, th, au, my, uk
WEBULL_REGION=us
# Optional. If set, account tools use this account when you omit account_id.
# Leave blank to auto-resolve the first account from list_accounts.
WEBULL_DEFAULT_ACCOUNT_ID=
# Leave this false unless you have read the trading section below.
WEBULL_ENABLE_TRADING=false
.env is gitignored. Never commit it, and never paste its contents anywhere.
4. Register with your MCP client
Edit .mcp.json and replace the placeholder paths with the absolute path to
your clone:
{
"mcpServers": {
"webull": {
"command": "/absolute/path/to/webull-mcp/.venv/bin/python",
"args": ["/absolute/path/to/webull-mcp/src/server.py"]
}
}
}
For Claude Code, either drop .mcp.json into a project directory where it will
be auto-detected, or register it explicitly:
claude mcp add webull -- /absolute/path/to/webull-mcp/.venv/bin/python /absolute/path/to/webull-mcp/src/server.py
Restart your client. Verify the server starts:
.venv/bin/python src/server.py
It should start and wait on stdio — no output is correct. Ctrl-C to exit.
On the first call that touches your account, the SDK performs a 2FA handshake
and caches a token under conf/. That directory is gitignored.
Enabling live trading (optional)
Only do this if you accept that every order is real. There is no sandbox, no paper account, and no undo.
- Set
WEBULL_ENABLE_TRADING=truein.env— the string must be exactlytrue. - Restart your MCP client. The trading tools are registered at import time, so a restart is required.
- Confirm with the
trading_statustool.
Two independent safety layers remain in place:
- Registration gate — with the flag unset or
false,place_order,preview_order, andcancel_orderare never exposed to the model at all. It cannot call a tool it cannot see. - Confirmation gate — even when enabled,
place_orderraises unless it is called withconfirm=True. Runpreview_orderfirst.
To disable again, set the flag back to false and restart.
Notes and limitations
- Region defaults to
us. SetWEBULL_REGIONfor other markets. - Order tools cover equities only — options trading is not implemented. Option support here is market data only.
- Webull exposes no option-chain listing endpoint. Build OCC symbols yourself.
Expired contracts return
INVALID_SYMBOL, and Friday expirations shift to Thursday when Friday is a market holiday. - Built on the unified, symbol-based
webull-openapi-python-sdk. The older splitwebull-python-sdk-*packages use a different, instrument-id-based API — do not install both. - The SDK writes logs under
logs/. These logs contain your App Key, your 2FA token, and your account IDs. They are gitignored; never attach one to a bug report without redacting it first.
Contributing
See CONTRIBUTING.md. Bug reports and PRs are welcome.
License
MIT.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.